tag:theconversation.com,2011:/fr/topics/artificial-intelligence-ai-90/articlesArtificial intelligence (AI) – The Conversation2024-03-26T16:03:37Ztag:theconversation.com,2011:article/2248822024-03-26T16:03:37Z2024-03-26T16:03:37ZWe built an AI tool to help set priorities for conservation in Madagascar: what we found<p><em>Artificial Intelligence (AI) – models that process large and diverse datasets and make predictions from them – can have many uses in nature conservation, such as remote monitoring (like the use of camera traps to study animals or plants) or data analysis. Some of these are controversial because AI can be trained to be biased, but others are valuable research tools.</em></p>
<p><em>Biologist Daniele Silvestro has developed an <a href="https://www.nature.com/articles/s41893-022-00851-6">AI tool</a> that can help identify conservation and restoration priorities. We asked him to tell us more about how it works and what it offers.</em></p>
<hr>
<h2>How does your artificial intelligence tool for conservation work?</h2>
<p>Artificial intelligence (AI) is a term indicating a broad family of models used to process large and diverse datasets and make predictions from them. </p>
<p>We built a model using biodiversity datasets as well as socioeconomic data. The aim was to identify optimal strategies to conserve nature. Our AI tool, Conservation Area Prioritisation through Artificial Intelligence (Captain), uses a type of AI called <a href="https://online.york.ac.uk/what-is-reinforcement-learning/">reinforcement learning</a>. This is a family of algorithms that optimises decisions within a dynamic environment. </p>
<p>The tool we built was the result of years of work involving an international team with experience in biology, sustainable economics, maths and computer science.</p>
<p>The software we developed can take multiple types of data as input, including biodiversity maps, species ranges, climate and predicted climate change, as well as socioeconomic data such as cost of land and a budget available for conservation action. It then processes this information and, based on a set conservation target (for example, to include all endangered species in a protected area, or to protect as many species as possible) it suggests a conservation policy.</p>
<p>The tool’s environment is a simulation of biodiversity, an artificial world with species and individuals that reproduce, migrate and die through time. We use the tool to look for the most appropriate conservation policy. </p>
<p>It works similarly to a video game where the player (called the agent) is the “brains” of our software. The goal of the game is to protect biodiversity and prevent as many species as possible from going extinct within a simulated environment that includes human pressure and climate change. </p>
<p>The agent observes the environment and tries to place protected areas in this environment in the best way. At the end of the game the agent gets a reward for each species it manages to save from extinction. It will have to play the game many times to learn how to best interpret the environment and place the protected areas. After that, the model is trained and can be used with real biodiversity data to identify conservation priorities that should maximise biodiversity protection. </p>
<h2>Why did you test the tool in Madagascar? What did you find?</h2>
<p>The <a href="https://www.kew.org/sites/default/files/2023-10/State%20of%20the%20World%27s%20Plants%20and%20Fungi%202023.pdf">State of the World’s Plants and Fungi report</a> showed that biodiversity is facing unprecedented threats, with as many as 45% of all plant species at risk of extinction. Together with climate change, this is one of the major challenges humanity faces, given our dependence on the natural world for our survival. </p>
<p>In a recent <a href="https://www.science.org/doi/full/10.1126/science.adf1466">paper</a> we summarised the extent of Madagascar’s extraordinary concentration of biodiversity with thousands of species of plants, animals and fungi. The project was led by Hélène Ralimanana of the Royal Botanic Gardens, Kew and Kew Madagascar Conservation Centre. </p>
<p>By applying the Captain tool to a dataset of endemic trees of Madagascar we were able to identify the most important areas for biodiversity protection in the country, for instance the area in the Sava region, where the Marojejy National Park has long been established. </p>
<p>Madagascar already has number of conservation areas and programmes. What our experiment shows is that the technology we developed can be used with real-world data. We hope it can guide conservation planning.</p>
<h2>Who do you think can use the Captain AI?</h2>
<p>We think it can help policy makers, practitioners and companies in guiding conservation and restoration planning. In particular, the software can use diverse types of data in addition to biodiversity data. For instance it can use costs and opportunity costs related to setting up protected or restoration areas. It can also use future climate scenarios. </p>
<h2>Is technology alone enough to conserve biodiversity?</h2>
<p>Certainly not. Technology can help us by crunching the numbers and disentangling complex data. But there are many aspects of conservation that are not easily quantifiable as numbers. There are aspects of cultural value of land and nature, and social and political issues related to the fair distribution of resources. These are issues for real humans to take into account, rather than artificial intelligence programs. </p>
<p>Technology and science can (and should) assist us in making decisions, but ultimately the protection and conservation of the natural world is and must be in the hands of humans, not software.</p><img src="https://counter.theconversation.com/content/224882/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Daniele Silvestro is a computational biologist at the University of Fribourg (Switzerland) and University of Gothenburg (Sweden). He is also a co-founder of CAPTAIN Technologies LTD.
D.S. acknowledges funding from the Swiss National Science Foundation (PCEFP3_187012), the Swedish Research Council (2019-04739), and the Swedish Foundation for Strategic Environmental Research MISTRA within the framework of the research programme BIOPATH (F 2022/1448).</span></em></p>Conservation of biodiversity is in the hands of humans but artificial intelligence can help guide decisions.Daniele Silvestro, Assistant Professor, Department of Biology, University of FribourgLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2238092024-03-26T12:39:43Z2024-03-26T12:39:43ZHow AI and a popular card game can help engineers predict catastrophic failure – by finding the absence of a pattern<figure><img src="https://images.theconversation.com/files/584159/original/file-20240325-10630-wq22k6.png?ixlib=rb-1.1.0&rect=84%2C498%2C1343%2C882&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Can you find a matching set?</span> <span class="attribution"><a class="source" href="https://en.wikipedia.org/wiki/File:Set_isomorphic_cards.svg">Cmglee/Wikimedia Commons</a>, <a class="license" href="http://creativecommons.org/licenses/by-sa/4.0/">CC BY-SA</a></span></figcaption></figure><p>Humans are very good at <a href="https://doi.org/10.1016/j.neuron.2018.05.013">spotting patterns</a>, or repeating features people can recognize. For instance, ancient Polynesians navigated across the Pacific by <a href="https://nla.gov.au/nla.obj-318911458/view?partId=nla.obj-318923632#page/n90/mode/1up">recognizing many patterns</a>, from the stars’ constellations to more subtle ones such as the directions and sizes of ocean swells.</p>
<p>Very recently, <a href="https://scholar.google.com/citations?user=L1lKOGsAAAAJ&hl=en">mathematicians like me</a> have started to study large collections of objects that have no patterns of a particular sort. How large can collections be before a specified pattern has to appear somewhere in the collection? Understanding such scenarios can have significant real-world implications: For example, what’s the smallest number of server failures that would lead to the severing of the internet?</p>
<p>Research from mathematician <a href="https://scholar.google.com/citations?user=b7P6YbkAAAAJ&hl=en">Jordan Ellenberg</a> at the University of Wisconsin and researchers at <a href="https://deepmind.google/">Google’s Deep Mind</a> have proposed a novel approach to this problem. Their work <a href="https://doi.org/10.1038/s41586-023-06924-6">uses artificial intelligence to find</a> large collections that don’t contain a specified pattern, which can help us understand some worst-case scenarios.</p>
<h2>Patterns in the card game Set</h2>
<p>The idea of patternless collections can be illustrated by a popular card game <a href="https://brilliant.org/wiki/set-game/">called Set</a>. In this game, players lay out 12 cards, face up. Each card has a different simple picture on it. They vary in terms of number, color, shape and shading. Each of these four features can have one of three values.</p>
<p>Players race to look for “sets,” which are groups of three cards in which every feature is either the same or different in each card. For instance, cards with one solid red diamond, two solid green diamonds and three solid purple diamonds form a set: All three have different numbers (one, two, three), the same shading (solid), different colors (red, green, purple) and the same shape (diamond).</p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/7AeEr9QtDF0?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
<figcaption><span class="caption">Marsha Falco originally created the game Set to help explain her research on population genetics.</span></figcaption>
</figure>
<p>Finding a set is usually possible – but not always. If none of the players can find a set from the 12 cards on the table, then they flip over three more cards. But they still might not be able to find a set in these 15 cards. The players continue to flip over cards, three at a time, until someone spots a set.</p>
<p>So what is the maximum number of cards you can lay out without forming a set?</p>
<p>In 1971, mathematician Giuseppe Pellegrino showed that the <a href="https://www.quantamagazine.org/set-proof-stuns-mathematicians-20160531/">largest collection of cards without a set is 20</a>. But if you chose 20 cards at random, “no set” would happen only <a href="https://www-cs-faculty.stanford.edu/%7Eknuth/programs/setset-all.w">about one in a trillion times</a>. And finding these “no set” collections is an extremely hard problem to solve.</p>
<h2>Finding ‘no set’ with AI</h2>
<p>If you wanted to find the smallest collection of cards with no set, you could in principle do an exhaustive search of every possible collection of cards chosen from the deck of 81 cards. But there are an enormous number of possibilities – on the order of 10<sup>24</sup> (that’s a “1” followed by 24 zeros). And if you increase the number of features of the cards from four to, say, eight, the complexity of the problem would overwhelm any computer doing an exhaustive search for “no set” collections.</p>
<p>Mathematicians love to think about computationally difficult problems like this. These complex problems, if approached in the right way, can become tractable. </p>
<p>It’s easier to find best-case scenarios – here, that would mean the fewest number of cards that could contain a set. But there were few known strategies that could explore bad scenarios – here, that would mean a large collection of cards that do not contain a set.</p>
<p>Ellenberg and his collaborators approached the bad scenario with a type of AI called <a href="https://theconversation.com/ai-to-z-all-the-terms-you-need-to-know-to-keep-up-in-the-ai-hype-age-203917">large language models, or LLMs</a>. The researchers first wrote computer programs that generate some examples of collections of many that contain no set. These collections typically have “cards” with more than four features.</p>
<p>Then they fed these programs to the LLM, which soon learned how to write many similar programs and choose the ones that give rise to the largest set-free collections to undergo the process again. Iterating that process by repeatedly tweaking the most successful programs enables them to find larger and larger set-free collections.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/584175/original/file-20240325-28-3w7tk3.png?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="Square of nine circles, four of which are colored blue, connected by grey, red, green, and yellow lines" src="https://images.theconversation.com/files/584175/original/file-20240325-28-3w7tk3.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/584175/original/file-20240325-28-3w7tk3.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=600&fit=crop&dpr=1 600w, https://images.theconversation.com/files/584175/original/file-20240325-28-3w7tk3.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=600&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/584175/original/file-20240325-28-3w7tk3.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=600&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/584175/original/file-20240325-28-3w7tk3.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=754&fit=crop&dpr=1 754w, https://images.theconversation.com/files/584175/original/file-20240325-28-3w7tk3.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=754&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/584175/original/file-20240325-28-3w7tk3.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=754&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">This is another version of a ‘no set,’ where no three components of a set are linked by a line.</span>
<span class="attribution"><a class="source" href="https://doi.org/10.1038/s41586-023-06924-6">Romera-Peredes et al./Nature</a>, <a class="license" href="http://creativecommons.org/licenses/by-sa/4.0/">CC BY-SA</a></span>
</figcaption>
</figure>
<p>This method allows people to explore disordered collections – in this instance, <a href="https://doi.org/10.1038/s41586-023-06924-6">collections of cards that contain no set</a> – in an entirely new way. It does not guarantee that researchers will find the absolute worst-case scenario, but they will find scenarios that are much worse than a random generation would yield.</p>
<p>Their work can help researchers understand how events might align in a way that leads to catastrophic failure. </p>
<p>For example, how vulnerable is the electrical grid to a malicious attacker who destroys select substations? Suppose that a bad collection of substations is one where they don’t form a connected grid. The worst-case scenario is now a very large number of substations that, when taken all together, still don’t yield a connected grid. The amount of substations excluded from this collection make up the smallest number a malicious actor needs to destroy to deliberately disconnect the grid.</p>
<p>The work of Ellenberg and his collaborators demonstrates yet another way that AI is a very powerful tool. But to solve very complex problems, at least for now, it still needs human ingenuity to guide it.</p><img src="https://counter.theconversation.com/content/223809/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>John Edward McCarthy is partially supported by National Science Foundation Grant
DMS 2054199. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.</span></em></p>What mathematicians call ‘disordered collections’ can help engineers explore real-world worst-case scenarios. The simple card game Set illustrates how to predict internet and electrical grid failures.John Edward McCarthy, Professor of Mathematics, Arts & Sciences at Washington University in St. LouisLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2258372024-03-25T16:37:51Z2024-03-25T16:37:51ZDating apps: Lack of regulation, oversight and competition affects quality, and millions stand to lose<figure><img src="https://images.theconversation.com/files/583280/original/file-20240320-24-xk2kwq.jpg?ixlib=rb-1.1.0&rect=0%2C19%2C6498%2C4299&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Dating apps have helped people make millions of connections.</span> <span class="attribution"><span class="source">(Shutterstock)</span></span></figcaption></figure><p>When Aleksandr Zhadan <a href="https://gizmodo.com/guy-used-chatgpt-talk-5-000-women-tinder-met-his-wife-1851228179">used ChatGPT</a> to talk to over 5,000 women on Tinder, it was a sign of things to come. </p>
<p>As artificial intelligence becomes more sophisticated and easily available, online dating is facing an onslaught of AI-powered fraud. The industry, which is dominated by a small number of incumbents, has already proven slow to respond to long-standing problems on its apps. AI will be its moment of reckoning — there are even apps that can <a href="https://www.cbc.ca/news/canada/artificial-intelligence-relationships-1.7148866">help people write their messages</a>.</p>
<p>Opponents of dating apps <a href="https://www.nytimes.com/2024/03/16/opinion/dating-apps-hinge-tinder-bumble.html">may be happy</a> to see the industry crash and burn. The rest of us should worry. Online dating plays an important, and I believe positive, role in our lives. It has made it easier for people to find relationships, and easier to find people with whom we are truly compatible.</p>
<p>As the industry careens towards disaster, regulators should be prepared to intervene.</p>
<p><div data-react-class="InstagramEmbed" data-react-props="{"url":"https://www.instagram.com/p/C3EHlNMt65J","accessToken":"127105130696839|b4b75090c9688d81dfd245afe6052f20"}"></div></p>
<h2>Real versus fake connections</h2>
<p>Zhadan’s case shows one of the challenges AI poses for online dating. Now, when we chat with someone on one of the apps, we cannot know if their answers are written by a chatbot, nor can we know how many other people they are talking to simultaneously. We also can’t know if someone’s photos have been <a href="https://mashable.com/article/using-ai-photo-generator-apps-for-dating-profile">produced with the help of an AI image generator</a> </p>
<p>But at least Zhadan was actually looking for love. Since the launch of ChatGPT in late 2022, the amount of outright fraud on dating apps, much of it powered by AI, has skyrocketed. According to cybersecurity company Arkose Labs, there was, between January 2023 and January 2024, <a href="https://www.arkoselabs.com/latest-news/how-criminals-are-manipulating-ai-to-target-dating-apps/">a staggering 2,000 per cent increase</a> in bot attacks on dating sites. </p>
<p>And this is just the beginning. AI is getting more powerful, and more convincingly human, all the time.</p>
<p>Even before AI appeared on the scene, fraud on dating apps was already a serious problem. Sign up for one of them and you’ll instantly find your feed clogged with an endless number of fake profiles. Most of them have been created for a specific purpose, which is to steal your money. Unfortunately, it works. </p>
<p>In 2023, 64,000 people in the United States admitted to being the victims of romance scams, most of which happen through dating apps — we can assume this is only a small portion of the actual cases. </p>
<p>The Federal Trade Commission <a href="https://www.ftc.gov/business-guidance/blog/2024/02/love-stinks-when-scammer-involved">measures the losses</a> for the year at US$1.14 billion. This has been going on for years, and the app companies have done little to stop it.</p>
<h2>Online connections, offline threats</h2>
<p>Fraud is not the only challenge faced by dating app users. <a href="https://www.kaspersky.com/about/press-releases/2024_nearly-a-quarter-of-online-daters-experience-digital-stalking">A quarter of them</a>, mostly women, have been stalked by someone they met online. Even more tragic are the cases of people being <a href="https://co.usembassy.gov/security-alert-risks-of-using-online-dating-applications/">assaulted or murdered</a>.</p>
<p>There are other issues: prices on the apps <a href="https://www.ft.com/content/83cd07a3-134c-4df7-ab6a-08752c724bbe">have gone up steadily</a> and innovation has come to a grinding halt. Ever since Tinder introduced <a href="https://techcrunch.com/2016/09/21/tinders-newest-app-tinder-stacks-lets-you-swipe-on-anything/">the card stack in 2016</a>, the design of the apps has hardly changed. </p>
<p>You swipe, match, message and hope for the best. It should perhaps be no surprise that customers are <a href="https://www.cnn.com/2024/02/14/business/dating-apps-2024-hinge-tinder-dg/index.html">getting fed up</a>.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/583458/original/file-20240321-30-md695s.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="Man using dating app on mobile phone" src="https://images.theconversation.com/files/583458/original/file-20240321-30-md695s.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/583458/original/file-20240321-30-md695s.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/583458/original/file-20240321-30-md695s.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/583458/original/file-20240321-30-md695s.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/583458/original/file-20240321-30-md695s.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/583458/original/file-20240321-30-md695s.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/583458/original/file-20240321-30-md695s.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">Online dating plays an important and positive role in people’s lives.</span>
<span class="attribution"><span class="source">(Shutterstock)</span></span>
</figcaption>
</figure>
<h2>Benefits to society</h2>
<p>While online dating certainly has its share of <a href="https://www.hachettebookgroup.com/titles/nancy-jo-sales/nothing-personal/9780316492799/">long-standing critics</a>, I have argued that, on balance, the apps are a <a href="https://www.taylorfrancis.com/chapters/mono/10.4324/9781315448848-7/sex-technology-neil-mcarthur">benefit to users and to society</a>. They are an efficient way to find partners, get us out of our social bubbles and encourage connections across class and race. </p>
<p>Precisely because of the important role the technology plays in our lives, we should pay attention to how the industry operates. The dating app companies are <a href="https://www.cnet.com/tech/services-and-software/what-you-need-to-know-about-tinders-new-verification-process/">finally starting</a> to do something to protect users. </p>
<p>But given how long fraud has plagued these apps, their response has been slow and pretty underwhelming. They need, at a minimum, better tools to detect fake accounts and remove them quickly. There is a lot more they could do as well. </p>
<p>They could require background checks for users, which <a href="https://www.pewresearch.org/short-reads/2023/02/02/key-findings-about-online-dating-in-the-u-s/">polls show</a> a majority of people support. They could put AI to use themselves, to flag signs of fraud during people’s private chats. And dating app companies could implement safety features to protect users when they meet in person, for instance making it easier to share with your friends or family the profiles of people you are meeting up with.</p>
<h2>Dominant players</h2>
<p>One explanation for the companies’ sluggish response will be familiar to any observer of big tech: the concentration of ownership. The dominant player, Match Group, owns <a href="https://faq.lert.matchgroup.com/en/brands">over 40 different apps</a>, including most of the well-known: Tinder, Match.com, OkCupid, Hinge and Plenty of Fish. Its only serious competitor for market share is <a href="https://ir.bumble.com/news/news-details/2022/Bumble-Inc.-acquires-popular-Gen-Z-dating-app-Fruitz/default.aspx">Bumble, which also owns Badoo and Fruitz</a>. </p>
<p>In the United States, Match Group and Bumble control <a href="https://www.start.io/blog/these-6-apps-own-85-of-the-mobile-dating-market-on-valentines-day-2024/">over three-quarters</a> of the <a href="https://www.businessofapps.com/data/dating-app-market/">market</a>. </p>
<p>Anti-trust authorities have never given the industry any serious scrutiny. Presumably, they do not think online dating is important enough to deserve it. But these companies have a lot of control over one of the most intimate aspects of our lives.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/583457/original/file-20240321-28-aq0kfb.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="a woman's hand holding a phone displaying a yellow background with the word BUMBLE" src="https://images.theconversation.com/files/583457/original/file-20240321-28-aq0kfb.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/583457/original/file-20240321-28-aq0kfb.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/583457/original/file-20240321-28-aq0kfb.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/583457/original/file-20240321-28-aq0kfb.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/583457/original/file-20240321-28-aq0kfb.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/583457/original/file-20240321-28-aq0kfb.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/583457/original/file-20240321-28-aq0kfb.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">In the United States, Match Group and Bumble control over three-quarters of the dating apps market.</span>
<span class="attribution"><span class="source">(Good Faces Agency/Unsplash)</span></span>
</figcaption>
</figure>
<p>Thirty per cent of all adults in the U.S., and over half of people under 30, have <a href="https://www.pewresearch.org/short-reads/2023/02/02/key-findings-about-online-dating-in-the-u-s/">used a dating app at some point</a>. One in 10 Americans is currently in a relationship with someone they met online. </p>
<p>The costs of fraud and abuse, in both human and financial terms, are huge. And the anti-competitive pressures in the industry are strong, given the network effect built into online dating: we want to be on the apps that everyone else is on.</p>
<p>Regulators should finally get involved. They should hold the companies accountable for fraud and abuse on their apps in order to force them to innovate to protect users. They should look closely at the prices they charge customers for premium features. The ultimate solution may be to break up the sector’s dominant players, Match Group and Bumble, in order to create real competition.</p>
<p>The inventors of dating apps deserve credit for enabling millions of connections that would never have happened otherwise. But if things don’t change, the companies could be in trouble and millions of people could be lonelier as a result.</p><img src="https://counter.theconversation.com/content/225837/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Neil McArthur does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.</span></em></p>Dating apps provide a valuable social service. The industry should be regulated to protect consumers, increase competition and address fraud.Neil McArthur, Director, Centre for Professional and Applied Ethics, University of ManitobaLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2257602024-03-22T12:32:33Z2024-03-22T12:32:33ZGenerative AI could leave users holding the bag for copyright violations<figure><img src="https://images.theconversation.com/files/583567/original/file-20240321-21-h56ai6.jpg?ixlib=rb-1.1.0&rect=0%2C0%2C5459%2C3612&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">How can users of AI tools like OpenAI's Sora video generator be sure they aren't producing copyright-violating content?</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/news-photo/in-this-photo-illustration-a-video-created-by-open-ais-news-photo/2008187952">Drew Angerer/AFP via Getty Images</a></span></figcaption></figure><p>Generative artificial intelligence has been hailed for its <a href="https://doi.org/10.1038/s41562-023-01751-1">potential to transform creativity</a>, and especially by lowering the <a href="https://doi.org/10.1145/3582269.3615596">barriers to content creation</a>. While the <a href="https://doi.org/10.1126/science.adh4451">creative potential of generative AI tools</a> has often been highlighted, the popularity of these tools poses questions about intellectual property and copyright protection. </p>
<p>Generative AI tools such as ChatGPT are powered by <a href="https://research.ibm.com/blog/what-are-foundation-models">foundational AI models</a>, or AI models <a href="https://hai.stanford.edu/news/what-foundation-model-explainer-non-experts">trained on vast quantities of data</a>. Generative AI is <a href="https://news.mit.edu/2023/explained-generative-ai-1109">trained on</a> billions of pieces of data taken from text or images scraped from the internet. </p>
<p>Generative AI uses very powerful machine learning methods such as <a href="https://doi.org/10.1038/nature14539">deep learning</a> and <a href="https://www.datacamp.com/tutorial/transfer-learning">transfer learning</a> on such vast repositories of data to understand the relationships among those pieces of data – for instance, which words tend to follow other words. This allows generative AI to perform a broad range of tasks that can <a href="https://doi.org/10.1145/3442188.3445922">mimic cognition and reasoning</a>. </p>
<p>One problem is that output from an AI tool can be <a href="https://crsreports.congress.gov/product/pdf/LSB/LSB10922">very similar to copyright-protected materials</a>. Leaving aside how generative models are trained, the challenge that widespread use of generative AI poses is how individuals and companies could be held liable when generative AI outputs infringe on copyright protections.</p>
<h2>When prompts result in copyright violations</h2>
<p><a href="https://doi.org/10.1017/dap.2022.10">Researchers</a> and <a href="https://www.tomshardware.com/tech-industry/artificial-intelligence/ai-image-generators-output-copyrighted-characters">journalists</a> have raised the possibility that through selective prompting strategies, people can end up creating text, images or video that violates copyright law. Typically, generative AI tools output an image, text or video but <a href="https://spectrum.ieee.org/midjourney-copyright">do not provide any warning about potential infringement</a>. This raises the question of how to ensure that users of generative AI tools do not unknowingly end up infringing copyright protection. </p>
<p>The legal argument advanced by generative AI companies is that AI trained on copyrighted works is not an infringement of copyright <a href="https://doi.org/10.1126/science.adi0656">since these models are not copying the training data</a>; rather, they are designed to learn the associations between the elements of writings and images like words and pixels. AI companies, including Stability AI, maker of image generator Stable Diffusion, contend that output images provided in response to a particular text prompt <a href="https://www.hollywoodreporter.com/business/business-news/artists-copyright-infringement-case-ai-art-generators-1235632929/">is not likely to be a close match</a> for any specific image in the training data. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/583571/original/file-20240321-16-98fgbv.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="a pair of hands, one holding a paint brush, over a painting" src="https://images.theconversation.com/files/583571/original/file-20240321-16-98fgbv.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/583571/original/file-20240321-16-98fgbv.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/583571/original/file-20240321-16-98fgbv.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/583571/original/file-20240321-16-98fgbv.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/583571/original/file-20240321-16-98fgbv.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/583571/original/file-20240321-16-98fgbv.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/583571/original/file-20240321-16-98fgbv.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">Some artists, including Kelly McKernan, who is shown here painting, have sued AI companies for copyright infringement.</span>
<span class="attribution"><a class="source" href="https://newsroom.ap.org/detail/AIVisualArtists/48ed9b3665824d72b59bf4b9a7151073/photo?Query=ai%20copyright&mediaType=photo&sortBy=creationdatetime:desc&dateRange=Anytime&totalCount=21&currentItemNo=1">AP Photo/George Walker IV</a></span>
</figcaption>
</figure>
<p>Builders of generative AI tools have argued that prompts do not reproduce the training data, which should protect them from claims of copyright violation. Some audit studies have shown, though, that <a href="https://www.tomshardware.com/tech-industry/artificial-intelligence/ai-image-generators-output-copyrighted-characters">end users of generative AI</a> can issue <a href="https://spectrum.ieee.org/midjourney-copyright">prompts that result in copyright violations</a> by producing works that <a href="https://garymarcus.substack.com/p/no-multimodal-chatgpt-is-not-going">closely resemble copyright-protected content</a>.</p>
<p>Establishing infringement requires <a href="https://houstonlawreview.org/article/92126-copyright-safety-for-generative-ai">detecting a close resemblance</a> between expressive elements of a stylistically similar work and original expression in particular works by that artist. Researchers have shown that methods such as <a href="http://dx.doi.org/10.48550/arXiv.2301.13188">training data extraction attacks</a>, which involve selective prompting strategies, and <a href="https://doi.org/10.48550/arXiv.2311.17035">extractable memorization</a>, which tricks generative AI systems into revealing training data, can recover individual training examples ranging from photographs of individuals to trademarked company logos. </p>
<p>Audit studies such as the one <a href="https://spectrum.ieee.org/midjourney-copyright">conducted by computer scientist Gary Marcus and artist Reid Southern</a> provide several examples where there can be little ambiguity about the degree to which visual generative AI models produce images that infringe on copyright protection. The New York Times provided a similar comparison of images showing how generative AI tools <a href="https://www.nytimes.com/interactive/2024/01/25/business/ai-image-generators-openai-microsoft-midjourney-copyright.html">can violate copyright protection</a>. </p>
<h2>How to build guardrails</h2>
<p>Legal scholars have dubbed the challenge in developing guardrails against copyright infringement into AI tools <a href="https://dx.doi.org/10.2139/ssrn.4438593">the “Snoopy problem</a>.” The more a copyrighted work is protecting a likeness – for example, the cartoon character Snoopy – the more likely it is a generative AI tool will copy it compared to copying a specific image. </p>
<p>Researchers in computer vision <a href="https://doi.org/10.1145/3466780">have long grappled with the issue</a> of how to detect copyright infringement, such as logos that are counterfeited or <a href="https://doi.org/10.1145/3466780">images that are protected by patents</a>. Researchers have also examined how <a href="https://doi.org/10.1145/3611309">logo detection can help identify counterfeit products</a>. These methods can be helpful in detecting violations of copyright. Methods to <a href="https://doi.org/10.1038/s42256-023-00733-2">establish content provenance and authenticity</a> could be helpful as well. </p>
<p>With respect to model training, AI researchers have suggested methods for making <a href="https://doi.org/10.1007/s42979-023-01767-4">generative AI models unlearn</a> <a href="https://openaccess.thecvf.com/content/CVPR2022W/HCIS/html/Kim_Efficient_Two-Stage_Model_Retraining_for_Machine_Unlearning_CVPRW_2022_paper.html">copyrighted data</a>. Some AI companies such as <a href="https://claudeai.uk/anthropic-says-no-client-data-used-in-ai-training/">Anthropic have announced pledges</a> to not use data produced by their customers to train advanced models such as Anthropic’s large language model Claude. Methods for AI safety such as <a href="https://doi.org/10.48550/arXiv.2403.04893">red teaming</a> – attempts to force AI tools to misbehave – or ensuring that the model training process <a href="https://doi.org/10.48550/arXiv.2302.10870">reduces the similarity</a> between the outputs of generative AI and copyrighted material may help as well.</p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/DUwDqSY8StE?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
<figcaption><span class="caption">Artists and technologists are fighting back against AI copyright infringement.</span></figcaption>
</figure>
<h2>Role for regulation</h2>
<p>Human creators know to decline requests to produce content that violates copyright. Can AI companies build similar guardrails into generative AI?</p>
<p>There’s no established approaches to build such guardrails into generative AI, nor are there any <a href="https://spectrum.ieee.org/midjourney-copyright">public tools or databases that users can consult</a> to establish copyright infringement. Even if tools like these were available, they could put an excessive burden on <a href="https://spectrum.ieee.org/midjourney-copyright">both users and content providers</a>.</p>
<p>Given that naive users can’t be expected to learn and follow best practices to avoid infringing copyrighted material, there are roles for policymakers and regulation. It may take a combination of legal and regulatory guidelines to ensure best practices for copyright safety. </p>
<p>For example, companies that build generative AI models could <a href="https://dx.doi.org/10.2139/ssrn.4438593">use filtering or restrict model outputs</a> to limit copyright infringement. Similarly, regulatory intervention may be necessary to ensure that builders of generative AI models <a href="https://dx.doi.org/10.2139/ssrn.4438593">build datasets and train models</a> in ways that reduce the risk that the output of their products infringe creators’ copyrights.</p><img src="https://counter.theconversation.com/content/225760/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Anjana Susarla receives funding from the National Institute of Health</span></em></p>With the right prompts, AI users can mimic copyrighted works. There’s no easy technical or legal fix.Anjana Susarla, Professor of Information Systems, Michigan State UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2195942024-03-21T18:58:32Z2024-03-21T18:58:32ZTreatments tailored to you: how AI will change NZ healthcare, and what we have to get right first<p>Imagine this: a novel virus is rapidly breaking out nationwide, resulting in an epidemic. The government introduces vaccination mandates and a choice of different vaccines is available. </p>
<p>But not everyone is getting the same vaccine. When you sign up for vaccination, you are sent a vial with instructions to send a sample of your saliva to the nearest laboratory. Just a few hours later you receive a message telling you which vaccine you should get. Your neighbour also signed up for vaccination. But their vaccine is different from yours. </p>
<p>Both of you are now vaccinated and protected, although each of you received your vaccines depending on “who you are”. Your genetics, age, gender, and myriad of other factors are captured in a “model” that predicts and determines the best option to protect you from the virus.</p>
<p>It all sounds a bit like science fiction. But since the <a href="https://www.genome.gov/human-genome-project">decoding of the human genome in 2003</a>, we have entered the age of precision prevention. </p>
<p>New Zealand has a long-standing newborn screening programme. This includes <a href="https://www.auckland.ac.nz/en/news/2023/11/30/newborn-genomic-sequencing-sick-babies.html">genome sequencing machines available nationwide</a> and a <a href="https://www.tewhatuora.govt.nz/our-health-system/genetic-health-service-nz/about/">genetic health service</a>. Programmes such as these open up the possibilities of public health genomics and precision public health for everyone.</p>
<p>The further expansion of these programmes, as well as the expansion of the use of artificial intelligence and machine learning to enable a shift to more personalised preventive care, will change how public health care is delivered.</p>
<p>At the same time, these developments raise wider concerns over individual choice versus the greater good, personal privacy, and who is responsible for the protection of New Zealanders and their health information.</p>
<p><div data-react-class="Tweet" data-react-props="{"tweetId":"1768051677905707508"}"></div></p>
<h2>What is precision prevention?</h2>
<p>Think of precision prevention (also known as personalised prevention) as public health action tailored to the individual rather than broader groups of society. </p>
<p>This targeted healthcare is achieved by balancing a range of variables (including your genes, life history and environment) with your risks (including everything that changes within you as you grow older). </p>
<p>While advances in genomics are making precision prevention possible, machine learning algorithms fuelled by our personal data have made it closer to a reality. </p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/its-2030-and-precision-medicine-has-changed-health-care-this-is-what-it-looks-like-90539">It's 2030, and precision medicine has changed health care – this is what it looks like</a>
</strong>
</em>
</p>
<hr>
<p>We generate data about ourselves every day – via social media, smartwatches and other wearable devices – helping to train algorithms to match medical prevention measures with individuals. </p>
<p>Combine all of these with AI-driven predictive modelling, and you have a system that can predict the current and future state of your health with an eerie level of accuracy, and <a href="https://www.forbes.com/sites/forbestechcouncil/2022/01/25/how-ai-could-predict-medical-conditions-and-revive-the-healthcare-system/?sh=362288726c47">help you take steps to prevent disease</a>. </p>
<h2>Safety and delay</h2>
<p>The Prime Minister’s Chief Science Advisor recently <a href="https://www.pmcsa.ac.nz/artificial-intelligence-2/ai-in-healthcare/">published a report</a> mapping out the landscape of artificial intelligence and machine learning in New Zealand over the next five years. </p>
<p>While the report authors didn’t specifically reference “precision prevention”, they did include examples of this approach, such as <a href="https://edition.cnn.com/2023/08/01/health/ai-breast-cancer-detection/index.html">computer vision augmented mammography</a>. </p>
<p>But as the report suggests, adoption tends to fall behind the pace of innovation in AI. Te Whatu Ora–Health New Zealand has also <a href="https://www.tewhatuora.govt.nz/our-health-system/digital-health/national-ai-and-algorithm-expert-advisory-group-naiaeag-te-whatu-ora-advice-on-the-use-of-large-language-models-and-generative-ai-in-healthcare/">not approved</a> emerging large language models and generative artificial intelligence tools as safe and effective for use in healthcare. </p>
<p>This means generative AI-driven precision prevention practices, such as conversational AI for public health messaging, may have to wait before they can be deemed safe to use. </p>
<h2>Move forward with caution</h2>
<p>There is much to be excited about the prospects of the use of artificial intelligence and machine learning in ushering in a new age of precision prevention and preventive health. But at the same time, we must temper this with caution. </p>
<p>Artificial intelligence and machine learning may increase access and utilisation of healthcare by lowering barriers to medical knowledge and reducing human bias. But government and medical agencies need to reduce barriers related to digital literacy and access to online platforms.</p>
<p>For those with limited access to online resources or who have limited digital literacy, the already existent inequity of access to care and health could worsen. </p>
<p>Artificial intelligence also has a <a href="https://arstechnica.com/gadgets/2023/04/generative-ai-is-cool-but-lets-not-forget-its-human-and-environmental-costs/">significant environmental impact</a>. <a href="https://arxiv.org/pdf/1906.02243.pdf">One study</a> found several common large AI models can emit over 270,000 tonnes of carbon dioxide during their life cycle.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/these-scientists-are-using-dna-to-target-new-drugs-for-your-genes-medicine-made-for-you-part-1-131986">These scientists are using DNA to target new drugs for your genes – Medicine made for you part 1</a>
</strong>
</em>
</p>
<hr>
<p>Finally, technology is a shifting landscape. Proponents of precision healthcare must be careful with children and marginalised communities and their access to resources. Maintaining privacy and choice is essential – everyone should be in a position to control what they share with the AI agents. </p>
<p>In the end, each of us is different, and we all have our different needs for our health and for our lives. Moving more people to preventive care through precision healthcare will reduce the financial burden on the health system. </p>
<p>But as the report from the prime minister’s chief science officer emphasises, machine learning algorithms are a nascent field. We need more public education and awareness before the technology becomes part of our everyday lives.</p><img src="https://counter.theconversation.com/content/219594/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Arindam Basu does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.</span></em></p>As New Zealand readies itself for AI-assisted medical treatment targeted to individuals, officials need to ensure the benefits outweigh the risks.Arindam Basu, Associate Professor, Epidemiology and Environmental Health, University of CanterburyLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2258542024-03-21T17:26:20Z2024-03-21T17:26:20ZAI’s excessive water consumption threatens to drown out its environmental contributions<p>Water is needed for development, production and consumption, yet we are overusing and polluting an unsubstitutable resource and system. </p>
<p>Eight safe and just boundaries for five domains (climate, biosphere, water, nutrients and aerosols) have been identified beyond which there is significant harm to humans and nature and the risk of crossing tipping points increases. Humans have already crossed the <a href="https://doi.org/10.1038/s41586-023-06083-8">safe and just Earth System Boundaries for water</a>. </p>
<p>To date, seven of the eight boundaries have been crossed, and although the aerosol boundary has not been crossed at the global level, it has been crossed at city level in many parts of the world.</p>
<p>For water, the safe and just boundaries specify that surface water flows should not fluctuate more than 20 per cent relative to the natural flow on a monthly basis; while groundwater withdrawal should not be more than the <a href="https://www.sciencedirect.com/topics/earth-and-planetary-sciences/groundwater-recharge">recharge rate</a>. Both of these boundaries have been crossed.</p>
<p>These thresholds have been crossed even though the minimum needs of the world’s poorest to access water and sanitation services <a href="https://www.unicef.org/wash#:%7E:text=Worldwide%2C%202.2%20billion%20people%20still,to%20handwashing%20facilities%20with%20soap">have not been met</a>. Addressing these needs will put an even greater pressure on <a href="https://doi.org/10.1038/s41893-022-00995-5">already-strained water systems</a>.</p>
<h2>AI’s potential</h2>
<p>Technological optimists argue that artificial intelligence (AI) holds the potential to solve the world’s water problems. Supporters of AI argue that it can help achieve both the environmental and social <a href="https://doi.org/10.1038/s41467-019-14108-y">Sustainable Development Goals</a> (SDGs), for example by designing systems to address shortages of teachers and doctors, increase crop yields and manage our energy needs.</p>
<p>In the past decade, research into this area has grown exponentially, with potential applications including increasing <a href="https://ieeexplore.ieee.org/document/8622984">water efficiency and monitoring in agriculture</a>, <a href="https://ieeexplore.ieee.org/document/10058801">water security</a> and <a href="https://doi.org/10.1016/j.psep.2019.11.014">enhancing wastewater treatment</a>. </p>
<p>AI-powered biosensors can more accurately <a href="https://doi.org/10.1016/j.gsd.2022.100888">detect toxic chemicals in drinking water</a> than current quality monitoring practices.</p>
<p>The potential for AI to change the water used in <a href="https://doi.org/10.1109/ACCESS.2022.3232485">agriculture</a> is evident through the building of smart machines, robots and sensors that optimize farming systems. </p>
<p>For example, <a href="https://doi.org/10.1109/ACCESS.2022.3232485">smart irrigation</a> automates irrigation through the collection and analysis of data to optimize water usage by <a href="https://www.sciencedirect.com/science/article/pii/S2772427122000791">improving efficiency</a> and <a href="http://article.sapub.org/10.5923.j.ijnc.20170701.01.html">detecting leakage</a>. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/583274/original/file-20240320-20-ot4d8m.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="rows of lettuce beds" src="https://images.theconversation.com/files/583274/original/file-20240320-20-ot4d8m.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/583274/original/file-20240320-20-ot4d8m.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=450&fit=crop&dpr=1 600w, https://images.theconversation.com/files/583274/original/file-20240320-20-ot4d8m.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=450&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/583274/original/file-20240320-20-ot4d8m.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=450&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/583274/original/file-20240320-20-ot4d8m.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=566&fit=crop&dpr=1 754w, https://images.theconversation.com/files/583274/original/file-20240320-20-ot4d8m.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=566&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/583274/original/file-20240320-20-ot4d8m.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=566&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">A smart irrigation system for green oak lettuce in Chiang Mai, Thailand.</span>
<span class="attribution"><span class="source">(Shutterstock)</span></span>
</figcaption>
</figure>
<p>As international development scholars who study the relationship between water, the environment and global inequality, we are curious about whether AI can actually make a difference or whether it exacerbates existing challenges. Although there is peer-reviewed literature on the use of AI for managing water and the SDGs, there are no peer-reviewed papers on the direct and indirect implications of AI on water use. </p>
<h2>AI and water use</h2>
<p>Initial research shows that AI has a significant water footprint. It uses water both for <a href="https://puiij.com/index.php/research/article/view/39/23">cooling the servers</a> that power its computations and for producing the energy it consumes. As AI becomes more integrated into our societies, its water footprint will inevitably grow. </p>
<p>The growth of ChatGPT and similar AI models has been hailed as “<a href="https://bootcamp.uxdesign.cc/is-chatgpt-the-new-google-5fdd0170c861">the new Google</a>.” But while a single Google search requires <a href="https://doi.org/10.1109/MSPEC.2010.5466789">half a millilitre of water in energy</a>, ChatGPT consumes <a href="https://doi.org/10.48550/arXiv.2304.03271">500 millilitres of water for every five to 50 prompts</a>. </p>
<p>AI <a href="https://puiij.com/index.php/research/article/view/39">uses</a> and <a href="https://doi.org/10.1016/j.jclepro.2014.08.061">pollutes</a> water through related hardware production. Producing the AI hardware involves resource-intensive mining for rare materials such as silicon, germanium, gallium, boron and phosphorous. Extracting these minerals has a <a href="https://doi.org/10.5897/JGRP2015.0495">significant impact on the environment and contributes to water pollution</a>. </p>
<p>Semiconductors and microchips require large volumes of water in the <a href="https://doi.org/10.1016/j.watcyc.2023.01.004">manufacturing stage</a>. Other hardware, such as for various <a href="https://doi.org/10.1021/acs.analchem.5b01653">sensors</a>, also have an associated water footprint.</p>
<p>Data centres provide the physical infrastructure for training and running AI, and their energy consumption <a href="https://www.iea.org/reports/electricity-2024">could double by 2026</a>. Technology firms using water to run and cool these data centres potentially require water withdrawals of 4.2 to 6.6 billion cubic metres by 2027.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/583276/original/file-20240320-30-2qnook.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="an aerial view of uniformly arranged rectangular buildings" src="https://images.theconversation.com/files/583276/original/file-20240320-30-2qnook.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/583276/original/file-20240320-30-2qnook.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=383&fit=crop&dpr=1 600w, https://images.theconversation.com/files/583276/original/file-20240320-30-2qnook.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=383&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/583276/original/file-20240320-30-2qnook.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=383&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/583276/original/file-20240320-30-2qnook.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=481&fit=crop&dpr=1 754w, https://images.theconversation.com/files/583276/original/file-20240320-30-2qnook.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=481&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/583276/original/file-20240320-30-2qnook.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=481&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">Microsoft data centers located in Noord-Holland, The Netherlands.</span>
<span class="attribution"><span class="source">(Shutterstock)</span></span>
</figcaption>
</figure>
<p>By comparison, <a href="https://sustainability.google/reports/google-2023-environmental-report/.">Google’s data centres</a> used over 21 billion litres of potable water in 2022, an increase of 20 per cent on its 2021 usage.</p>
<p>Training an AI at the computing level of a human brain for one year can cost <a href="https://doi.org/10.5281/zenodo.7855594">126,000 litres of water</a>. Each year the computing power needed to train AI <a href="https://openai.com/research/ai-and-compute">increases tenfold</a>, requiring more resources. </p>
<p>Water use of big tech companies’ data centres is grossly underestimated — for example, the <a href="http://www.aquatechtrade.com/news/industrial-water/microsoft-data-centre-uses-too-much-water">water consumption at Microsoft’s Dutch data centre was four times their initial plans</a>. Demand for water for cooling will only <a href="https://procido.com/2023/09/27/how-artificial-intelligence-ai-is-stealing-your-drinking-water/">increase</a> because of rising average temperatures due to climate change.</p>
<h2>Conflicting needs</h2>
<p>The technology sector’s water demand is so high that communities are protesting against it as it threatens their livelihoods. Google’s data centre in drought-prone The Dalles, Ore. is sparking concern as it uses a <a href="https://www.oregonlive.com/silicon-forest/2022/12/googles-water-use-is-soaring-in-the-dalles-records-show-with-two-more-data-centers-to-come.html">quarter of the city’s water</a>. </p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/9Kqfdq8ljUI?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
<figcaption><span class="caption">The Associated Press looks at Google’s water consumption in The Dalles, Ore.</span></figcaption>
</figure>
<p>Taiwan, responsible for 90 per cent of the world’s <a href="https://www.economist.com/special-report/2023/03/06/taiwans-dominance-of-the-chip-industry-makes-it-more-important">advanced semiconductor chip production</a>, has resorted to cloud seeding, water desalination, interbasin water transfers and halting irrigation for 180,000 hectares <a href="https://www.nytimes.com/2021/04/08/technology/taiwan-drought-tsmc-semiconductors.html">to address its water needs</a>. </p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/the-microchip-industry-would-implode-if-china-invaded-taiwan-and-it-would-affect-everyone-206335">The microchip industry would implode if China invaded Taiwan, and it would affect everyone</a>
</strong>
</em>
</p>
<hr>
<h2>Locating data centres</h2>
<p>As water becomes increasingly expensive and scarce in relation to demand, companies are now strategically placing their data centres in the <a href="https://news.mongabay.com/2023/11/the-cloud-vs-drought-water-hog-data-centers-threaten-latin-america-critics-say/">developing world</a> — even in dry sub-Saharan Africa, <a href="https://www.howwemadeitinafrica.com/africas-data-centre-boom/156344/">data centre investments are increasing</a>. </p>
<p>Google’s planned data centre in Uruguay, which recently suffered its <a href="https://hir.harvard.edu/running-dry-the-battle-for-water-security-in-uruguay-and-why-it-foreshadows-a-greater-issue/">worst drought in 74 years</a>, would require 7.6 million litres per day, <a href="https://www.theguardian.com/world/2023/jul/11/uruguay-drought-water-google-data-center">sparking widespread protest</a>. </p>
<p>What emerges is a familiar picture of geographic inequality, as developing countries find themselves caught in a dilemma between the economic benefits offered by international investment and the strain this places on local water resources availability. </p>
<p>We believe there is sufficient evidence for concern that the rapid uptake of AI risks exacerbating the water crises rather than help addressing them. As yet, there are no systematic studies on the AI industry and its water consumption. Technology companies have been tightlipped about the water footprint of their new products. </p>
<p>The broader question is: Will the social and environmental contributions of AI be overshadowed by its huge water footprint?</p><img src="https://counter.theconversation.com/content/225854/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Joyeeta Gupta receives funding from the Netherlands Enterprise Agency (RVO), under grant number 5000005700 and case number 31184622</span></em></p><p class="fine-print"><em><span>Hilmer Bosch receives funding from the Netherlands Enterprise Agency (RVO), under grant number 5000005700 and case number 31184622</span></em></p><p class="fine-print"><em><span>Luc van Vliet receives funding from the Netherlands Enterprise Agency (RVO), under grant number 5000005700 and case number 31184622</span></em></p>Artificial intelligence promises revolutionary solutions to global challenges, but the water costs to produce and power AI hardware and infrastructure may exceed the benefits.Joyeeta Gupta, Professor, Social and Behavioural Sciences, University of AmsterdamHilmer Bosch, Postdoctoral researcher on the Global Commission on the Economics of Water, University of AmsterdamLuc van Vliet, Researcher, Human Geography, University of AmsterdamLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2252162024-03-21T09:29:42Z2024-03-21T09:29:42Z‘Empathetic’ AI has more to do with psychopathy than emotional intelligence – but that doesn’t mean we can treat machines cruelly<p>AI has long since surpassed humans in cognitive matters that were once considered the supreme disciplines of human intelligence like chess or Go. Some even believe it is superior when it comes to human emotional skills such as empathy. This does not just seem to be some companies’ talking big for marketing reasons; empirical studies suggest that people perceive <a href="https://www.theguardian.com/technology/2023/apr/28/ai-has-better-bedside-manner-than-some-doctors-study-finds">ChatGPT in certain health situations as more empathic than human medical staff</a>. Does this mean that AI is really empathetic?</p>
<h2>A definition of empathy</h2>
<p>As a psychologically informed philosopher, I define genuine empathy according to <a href="https://www.taylorfrancis.com/chapters/oa-edit/10.4324/9781003333739-5/seeing-others-ends-catrin-misselhorn">three criteria</a>: </p>
<ul>
<li><p>Congruence of feelings: empathy requires that the person who empathizes to feels what it is like to experience the other’s emotions in a specific situation. This distinguishes empathy from a mere rational understanding of emotions. </p></li>
<li><p>Asymmetry: the person who feels empathy only has the emotion because another individual has it and it is more appropriate to the other’s situation than to their own. For this reason, empathy is not just a shared emotion like the shared joy of parents over the progress of their offspring, where the asymmetry-condition is not met. </p></li>
<li><p>Other-awareness: There must be at least a rudimentary awareness that empathy is about the feelings of another individual. This accounts for the difference between empathy and emotional contagion which occurs if one catches a feeling or an emotion like a cold. This happens, for instance, when kids start to cry when they see another kid crying.</p></li>
</ul>
<h2>Empathetic AI or psychopathic AI?</h2>
<p>Given this definition, it’s clear that artificial systems cannot feel empathy. They do not know what it’s like to feel something. This means that they cannot fulfil the congruence condition. Consequently, the question of whether what they feel corresponds to the asymmetry and other-awareness condition does not even arise. What artificial systems can do is recognise emotions, be it on the basis of facial expressions, vocal cues, physiological patterns or affective meanings; and they can simulate empathic behaviour by ways of speech or other modes of emotional expression. </p>
<p>Artificial systems hence show similarities to what common sense calls a psychopath: despite being unable to feel empathy, they are capable to recognize emotions on the basis of objective signs, to mimic empathy and to use this ability for manipulative purposes. Unlike psychopaths, artificial systems do not set these purposes by themselves, but are given them by their designers. So-called empathetic AI is often supposed to make us behave in a desired way, such as not getting upset when driving, learning with greater motivation, working more productively, buying a certain product – or voting for a certain political candidate. But then does not everything depend on how good the purposes are for which empathy-simulating AI is used?</p>
<h2>Empathy-simulating AI in the context of care and psychotherapy</h2>
<p>Take care and psychotherapy, which aim to nurture people’s well-being. You might think that the use of empathy-simulating AI in these areas is definitely a good thing. Would they not be wonderful care-givers and social companions for old people, <a href="https://bioedge.org/disability/sex-robots-can-offer-the-intimacy-that-the-elderly-disabled-crave-says-bioethicist/">loving partners for the disabled</a>, or <a href="https://www.theguardian.com/lifeandstyle/2024/mar/02/can-ai-chatbot-therapists-do-better-than-the-real-thing">perfect psychotherapists that have the benefit of being available 24/7</a>?</p>
<p>Such questions ultimately concern what it means to be a human being. Is it enough for a lonely, old or mentally disturbed person to project emotions onto an artefact devoid of feelings, or is it important for a person to experience recognition for themselves and their suffering in an interpersonal relationship? </p>
<h2>Respect or tech?</h2>
<p>From an ethical perspective, it is a matter of respect whether there is someone who empathically acknowledges the needs and the suffering of a person as such. By taking away recognition by another subject, the person in need of care, companionship or psychotherapy is treated as a mere object because ultimately this is based on the assumption that it does not matter whether anybody really listens to the person. They do not have a moral claim that their feelings, needs and suffering is perceived by someone who can really understand them.
Using <a href="https://www.hachettebookgroup.com/titles/evgeny-morozov/to-save-everything-click-here/9781610391399/?lens=publicaffairs">empathy-simulating AI in care and psychotherapy is ultimately another case of technological solutionism</a>, i.e., the naïve assumption that there is a technological fix for every problem, including loneliness and mental “malfunctions”. Outsourcing these issues to artificial systems prevents us from seeing the social causes for loneliness and mental disorders in the larger context of society.</p>
<p>In addition, designing artificial systems to appear as someone or something that has emotions and feels empathy would mean that such devices always have a manipulative character because they address very subliminal mechanisms of anthropomorphisation. This fact is used in commercial applications to get users to unlock a paid premium level: or customers pay with their data. Both practices are particularly problematic for <a href="https://theconversation.com/i-tried-the-replika-ai-companion-and-can-see-why-users-are-falling-hard-the-app-raises-serious-ethical-questions-200257">vulnerable groups</a>, which are at stake here. Even people who do not belong to vulnerable groups and are perfectly aware that an artificial system has no feelings will still react empathically to it as if it did.</p>
<h2>Empathy with artificial systems – all too human</h2>
<p>It is a well-studied phenomenon that humans react with empathy towards artificial systems that display certain human or animal-like characteristics. This process is largely based on perceptual mechanisms which are not consciously accessible. Perceiving a sign that another individual is undergoing a certain emotion produces a congruent emotion in the observer. Such a sign can be a typical behavioural manifestation of an emotion, a facial expression or an event that typically causes a certain emotion. Evidence from brain MRI scans shows that the same neural structures <a href="https://www.livescience.com/28947-humans-show-empathy-for-robots.html">get activated when humans feel empathy with robots</a>.</p>
<p>Although empathy might not be strictly necessary for morality, it plays an important moral role. For this reason, our empathy toward human-like (or animal-like) robots imposes at least indirect moral constraints on how we should treat these machines. It is morally wrong to habitually abuse robots that elicit empathy as doing so negatively affects our capacity to feel empathy, which is an <a href="https://link.springer.com/chapter/10.1007/978-3-658-37641-3_7">important source of moral judgment, motivation, and development</a>. </p>
<p>Does this mean that we have to establish a robot-rights league? That would be premature, as robots do not have moral claims by themselves. Empathy with robots is only indirectly morally relevant due to its effects on human morality. But we should carefully consider whether and in which areas we really want robots that simulate and evoke empathy in human beings as we run <a href="https://link.springer.com/chapter/10.1007/978-3-658-37641-3_10">the risk of distorting or even destroying our social practices if they became pervasive</a>.</p><img src="https://counter.theconversation.com/content/225216/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Catrin Misselhorn ne travaille pas, ne conseille pas, ne possède pas de parts, ne reçoit pas de fonds d'une organisation qui pourrait tirer profit de cet article, et n'a déclaré aucune autre affiliation que son organisme de recherche.</span></em></p>Artificial Intelligence’s capacity to mimic and identify emotions is worlds away from the human one to feel.Catrin Misselhorn, Professor of philosophy, Georg-August-Universität Göttingen Licensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2256372024-03-20T12:21:04Z2024-03-20T12:21:04ZAI can help predict whether a patient will respond to specific tuberculosis treatments, paving way for personalized care<figure><img src="https://images.theconversation.com/files/582664/original/file-20240318-30-l7zhd0.jpg?ixlib=rb-1.1.0&rect=0%2C0%2C2121%2C1412&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Tuberculosis typically infects the lungs but can spread to the rest of the body.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/pulmonary-tuberculosis-royalty-free-image/468719560">stockdevil/iStock via Getty Images Plus</a></span></figcaption></figure><p>Tuberculosis is the <a href="https://www.who.int/news-room/fact-sheets/detail/tuberculosis">world’s deadliest bacterial infection</a>. It afflicted over 10 million people and took 1.3 million lives in 2022. These numbers are predicted to increase dramatically because of the spread of multidrug-resistant TB. </p>
<p>Why does one TB patient recover from the infection while another succumbs? And why does one drug work in one patient but not another, even if they have the same disease?</p>
<p>People have been <a href="https://globaltb.njms.rutgers.edu/abouttb/historyoftb.php">battling TB for millennia</a>. For example, researchers have found Egyptian mummies from 2400 BCE that show signs of TB. While TB infections occur worldwide, the countries with the highest number of multidrug-resistant TB cases are <a href="https://doi.org/10.1016/S1473-3099(19)30568-7">Ukraine, Moldova, Belarus and Russia</a>. </p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/umfQyVT9Fg0?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
<figcaption><span class="caption">The COVID-19 pandemic set back progress in addressing many health conditions, including TB.</span></figcaption>
</figure>
<p>Researchers predict that the <a href="https://doi.org/10.1016/S1473-3099(22)00214-6">ongoing war in Ukraine</a> will result in an increase in multidrug-resistant TB cases because of health care disruptions. Additionally, the <a href="https://doi.org/10.1016%2FS2213-2600(21)00496-3">COVID-19 pandemic</a> reduced access to TB diagnosis and treatment, reversing decades of progress worldwide. </p>
<p>Rapidly and holistically analyzing available medical data can help optimize treatments for each patient and reduce drug resistance. In our recently published research, <a href="https://systemsbiologylab.org/">my team</a> <a href="https://scholar.google.com/citations?user=3xkdD2AAAAAJ&hl=en">and I</a> describe a new <a href="https://doi.org/10.1016/j.isci.2024.109025">AI tool</a> we developed that uses worldwide patient data to guide more personalized and effective treatment of TB.</p>
<h2>Predicting success or failure</h2>
<p>My team and I wanted to identify what variables can predict how a patient responds to TB treatment. So we analyzed more than 200 types of clinical test results, medical imaging and drug prescriptions from over 5,000 TB patients in 10 countries. We examined demographic information such as age and gender, prior treatment history and whether patients had other conditions. Finally, we also analyzed data on various TB strains, such as what drugs the pathogen is resistant to and what genetic mutations the pathogen had.</p>
<p>Looking at enormous datasets like these can be overwhelming. Even most existing AI tools have had difficulty analyzing large datasets. <a href="https://doi.org/10.1136/bmjopen-2020-044687">Prior studies</a> using AI have focused on a single data type – such as imaging or age alone – and had limited success predicting TB treatment outcomes. </p>
<p>We used an approach to AI that allowed us to analyze a large and diverse number of variables simultaneously and identify their relationship to TB outcomes. Our <a href="https://theconversation.com/what-is-a-black-box-a-computer-scientist-explains-what-it-means-when-the-inner-workings-of-ais-are-hidden-203888">AI model was transparent</a>, meaning we can see through its inner workings to identify the most meaningful clinical features. It was <a href="https://doi.org/10.1038/s41591-022-01981-2">also multimodal</a>, meaning it could interpret different types of data at the same time. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/582665/original/file-20240318-18-yertme.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="Microscopy image of rod-shaped TB bacteria stained green" src="https://images.theconversation.com/files/582665/original/file-20240318-18-yertme.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/582665/original/file-20240318-18-yertme.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=548&fit=crop&dpr=1 600w, https://images.theconversation.com/files/582665/original/file-20240318-18-yertme.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=548&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/582665/original/file-20240318-18-yertme.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=548&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/582665/original/file-20240318-18-yertme.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=689&fit=crop&dpr=1 754w, https://images.theconversation.com/files/582665/original/file-20240318-18-yertme.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=689&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/582665/original/file-20240318-18-yertme.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=689&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption"><em>Mycobacterium tuberculosis</em> spreads through aerosol droplets.</span>
<span class="attribution"><a class="source" href="https://www.flickr.com/photos/nihgov/52730545339/">NIAID/NIH via Flickr</a></span>
</figcaption>
</figure>
<p>Once we trained our AI model on the dataset, we found that it could <a href="https://doi.org/10.1016/j.isci.2024.109025">predict treatment prognosis with 83% accuracy</a> on newer, unseen patient data and outperform existing AI models. In other words, we could feed a new patient’s information into the model and the AI would determine whether a specific type of treatment will either succeed or fail.</p>
<p>We observed that clinical features <a href="https://doi.org/10.1016/j.isci.2024.109025">related to nutrition</a>, particularly lower BMI, are associated with treatment failure. This supports the use of interventions to improve nourishment, as TB is typically <a href="https://doi.org/10.1016%2Fj.jctube.2022.100311">more prevalent in undernourished populations</a>. </p>
<p>We also found that <a href="https://doi.org/10.1016/j.isci.2024.109025">certain drug combinations</a> worked better in patients with certain types of drug-resistant infections but not others, leading to treatment failure. <a href="https://doi.org/10.1016/j.drudis.2022.04.006">Combining drugs that are synergistic</a>, meaning they enhance each other’s potency in the lab, could result in better outcomes. Given the complex environment in the body compared with conditions in the lab, it has so far been unclear whether synergistic relationships between drugs in the lab hold up in the clinic. Our results suggest that <a href="https://doi.org/10.1128/mbio.02627-19">using AI to weed out antagonistic drugs</a>, or drugs that inhibit or counteract each other, early in the drug discovery process can avoid treatment failures down the line. </p>
<h2>Ending TB with the help of AI</h2>
<p>Our findings may help researchers and clinicians meet the World Health Organization’s goal to <a href="https://www.who.int/publications/i/item/WHO-HTM-TB-2015.19">end TB by 2035</a>, by highlighting the relative importance of different types of clinical data. This can help prioritize public health efforts to mitigate TB. </p>
<p>While the performance of our AI tool is promising, it isn’t perfect in every case, and more training is needed before it can be used in the clinic. Demographic diversity can be high within a country and may even vary between hospitals. We are working to make this tool more generalizable across regions. </p>
<p>Our goal is to eventually tailor our AI model to identify drug regimens suitable for individuals with certain conditions. Instead of a one-size-fits-all treatment approach, we hope that studying multiple types of data can help physicians personalize treatments for each patient to provide the best outcomes.</p><img src="https://counter.theconversation.com/content/225637/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Sriram Chandrasekaran receives funding from the US National Institutes of Health. </span></em></p>People have been battling tuberculosis for thousands of years, and drug-resistant strains are on the rise. Analyzing large datasets with AI can help humanity gain a crucial edge over the disease.Sriram Chandrasekaran, Associate Professor of Biomedical Engineering, University of MichiganLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2258942024-03-19T19:44:12Z2024-03-19T19:44:12ZCan AI improve football teams’ success from corner kicks? Liverpool and others are betting it can<figure><img src="https://images.theconversation.com/files/582686/original/file-20240318-26-ut2che.jpg?ixlib=rb-1.1.0&rect=0%2C0%2C4000%2C2005&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><span class="source">Google DeepMind</span></span></figcaption></figure><p>Last Sunday, Liverpool faced Manchester United in the <a href="https://www.espn.com.au/football/report/_/gameId/699283">quarter finals of the FA Cup</a> – and in the final minute of extra time, with the score tied at three-all, Liverpool had the crucial opportunity of a corner kick. A goal would surely mean victory, but losing possession could be risky.</p>
<p>What was Liverpool to do? Attack or play it safe? And if they were to attack, how best to do it? What kind of delivery, and where should players be waiting to attack the ball?</p>
<p>Set-piece decisions like this are vital not only in football but in many other competitive sports, and traditionally they are made by coaches on the basis of long experience and analysis. However, Liverpool has recently been looking to an unexpected source for advice: researchers at the Google-owned UK-based artificial intelligence (AI) lab <a href="https://deepmind.google/discover/blog/advancing-sports-analytics-through-ai-research/">DeepMind</a>.</p>
<p>In a <a href="https://www.nature.com/articles/s41467-024-45965-x">paper published today</a> in Nature Communications, DeepMind researchers describe an AI system for football tactics called TacticAI, which can assist in developing successful corner kick routines. The paper says experts at Liverpool favoured TacticAI’s advice over existing tactics in 90% of cases.</p>
<h2>What TacticAI can do</h2>
<p>At a corner kick, play stops and each team has the chance to organise its players on the field before the attacking team kicks the ball back into play – usually with a specific prearranged plan in mind that will (hopefully) let them score a goal. Advice on these prearranged plans or routines is what TacticAI sets out to offer.</p>
<p>The package has three components: one that predicts which player is most likely to receive the ball in a given scenario, another that predicts whether a shot on goal will be taken, and a third that recommends how to adjust the position of players to increase or decrease the chances of a shot on goal.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/582707/original/file-20240319-28-xag9u9.png?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="A diagram showing a soccer field with player positions marked, as well as a network diagram." src="https://images.theconversation.com/files/582707/original/file-20240319-28-xag9u9.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/582707/original/file-20240319-28-xag9u9.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=258&fit=crop&dpr=1 600w, https://images.theconversation.com/files/582707/original/file-20240319-28-xag9u9.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=258&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/582707/original/file-20240319-28-xag9u9.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=258&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/582707/original/file-20240319-28-xag9u9.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=325&fit=crop&dpr=1 754w, https://images.theconversation.com/files/582707/original/file-20240319-28-xag9u9.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=325&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/582707/original/file-20240319-28-xag9u9.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=325&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">TacticAI represents a corner-kick setup as a ‘graph’ of player positions and relationships, which it then uses to make predictions.</span>
<span class="attribution"><a class="source" href="https://doi.org/10.1038/s41467-024-45965-x">Wang et al. / Nature Communications</a></span>
</figcaption>
</figure>
<p>Trained on a dataset of 7,176 corner kicks from Premier League matches, TacticAI used a technique called “geometric deep learning” to identify key strategic patterns.</p>
<p>The researchers say this approach could be applied not only to football, but to any sport in which a stoppage in the game allows teams to deliberately manoeuvre players into place unopposed, and plan the next sequence of play. In football, it could also be expanded in future to incorporate throw-in routines as well as other set pieces such as attacking free kicks.</p>
<h2>Vast amounts of data</h2>
<p>AI in football is not new. Even in amateur and semi-professional football, AI-powered auto-tracking camera systems are becoming commonplace, for example. At the last men’s and women’s World Cups in 2022 and 2023, AI in conjunction with advanced ball-tracking technology produced semi-automated offside decisions with an unprecedented level of accuracy.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/games-by-numbers-machine-learning-is-changing-sport-38973">Games by numbers: machine learning is changing sport</a>
</strong>
</em>
</p>
<hr>
<p>Professional football clubs have analytical departments using AI at every level of the game, predominantly in the areas of <a href="https://www.wired.com/story/ai-football-soccer-scouting/">scouting</a>, <a href="https://www.engadget.com/will-ai-revolutionize-professional-soccer-recruitment-130045118.html">recruitment</a> and <a href="https://theathletic.com/4966509/2023/10/19/wearable-technology-in-football/">athlete monitoring</a>. Other research has also tried to <a href="https://www.mdpi.com/1424-8220/23/9/4506">predict players’ shots on goal</a>, or guess from a video what <a href="https://www.nature.com/articles/s41598-022-12547-0">off-screen players are doing</a>. </p>
<p>Bringing AI into tactical decisions promises to offer coaches a more objective and analytical approach to the game. Algorithms can process vast amounts of data, identifying patterns that may not be apparent to the naked eye, giving teams valuable insights into their own performance as well as that of their opponents. </p>
<h2>A useful tool</h2>
<p>AI may be a useful tool, but it cannot make decisions about match play alone. An algorithm might suggest the optimal positional setup for an in-swinging corner or how best to exploit the opposition’s defensive tactics. </p>
<p>What AI cannot do is make decisions on the fly – like deciding whether to take a corner quickly to exploit an opponent’s lapse in concentration. </p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/b1zjjf5EN1g?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
<figcaption><span class="caption">Sometimes the best move is a speedy reaction to conditions on the ground, not an elaborate prearranged set play.</span></figcaption>
</figure>
<p>There’s also something to be said for allowing players creative licence in some situations. Once teams are using AI to suggest the optimal corner strategy, opponents will doubtless counter with their own AI-prompted defensive setup.</p>
<p>So while the tech behind TacticAI is very interesting, it remains to be seen whether it can evolve to be useful in open play. Could AI get to the stage where it can recognise the best tactical player substitution in a given situation? </p>
<p>DeepMind researchers have advanced decision-making like this in their sights for <a href="https://dl.acm.org/doi/10.1613/jair.1.12505">future research</a>, but will it ever reach a point where coaches would trust it?</p>
<p>My sense from discussions with people in the industry is many believe AI should only be used as an input to decision-making, and not be allowed to make decisions itself. There is no substitute for the experience and instinct of the best coaches, the intangible ability to feel what the game needs, to make a change in formation, to play someone out of position. </p>
<h2>Smart tactics – but what about strategy?</h2>
<p>Coming back to that crucial Liverpool corner in last Sunday’s FA Cup quarter final: we don’t know whether Liverpool’s manager Jürgen Klopp considered AI advice, but the decision was made to play an attacking corner kick, presumably in the hope of scoring a last-minute winner. </p>
<p>The out-swinging delivery into the box may well have been the tactic with the highest probability of scoring a goal – but things rapidly went wrong. Manchester United gained possession of the ball, moved it down the pitch on the counterattack and slotted home the winning goal, sending Liverpool out of the tournament at the last moment.</p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/DKk8N2PYwCA?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
<figcaption><span class="caption">Even the best tactics can go wrong.</span></figcaption>
</figure>
<p>So while AI might suggest the optimal delivery and setup for a set piece, a coach might decide the wiser move is to play safe and avoid the risk of a counterattack. If TacticAI continues its career progression as a coaching assistant, it will no doubt learn that keeping the ball in the corner and playing for penalties may sometimes be the better option.</p><img src="https://counter.theconversation.com/content/225894/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Mark Scanlan does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.</span></em></p>A new AI system may improve soccer tactics in 90% of corner kicks – but is it ready for the big leagues?Mark Scanlan, Lecturer, Edith Cowan UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2247862024-03-19T18:17:34Z2024-03-19T18:17:34ZDeepfakes are still new, but 2024 could be the year they have an impact on elections<figure><img src="https://images.theconversation.com/files/580733/original/file-20240308-30-tf2e5r.jpg?ixlib=rb-1.1.0&rect=0%2C0%2C3865%2C2582&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/deep-fake-ai-face-swap-video-2376208005">Tero Vesalainen / Shutterstock</a></span></figcaption></figure><p>Disinformation caught many people off guard during the <a href="https://www.europarl.europa.eu/RegData/etudes/ATAG/2018/620230/EPRS_ATA(2018)620230_EN.pdf">2016 Brexit referendum</a> and <a href="https://www.nature.com/articles/s41467-018-07761-2">US presidential election</a>. Since then, a mini-industry has developed to analyse and counter it.</p>
<p>Yet despite that, we have entered 2024 – a year of <a href="https://en.wikipedia.org/wiki/List_of_elections_in_2024">more than 40 elections</a> worldwide – more fearful than ever about disinformation. In many ways, the problem is more challenging than it was in 2016. </p>
<p>Advances in technology since then are one reason for that, in particular the development that has taken place with synthetic media, otherwise known as deepfakes. It is increasingly difficult to know whether media has been fabricated by a computer or is based on something that really happened. </p>
<p>We’ve yet to really understand how big an impact deepfakes could have on elections. But a number of examples point the way to how they may be used. This may be the year when lots of mistakes are made and lessons learned.</p>
<p>Since the disinformation propagated around the votes in 2016, researchers have produced countless books and papers, journalists have retrained as <a href="https://www.poynter.org/fact-checking/2022/391-global-fact-checking-outlets-slow-growth-2022/">fact checking and verification experts</a>, governments have participated in <a href="https://www.igcd.org/">“grand committees”</a> and centres of excellence. Additionally, <a href="https://royalsociety.org/blog/2022/03/how-libraries-can-fight-disinformation/">libraries</a> have become the focus of resilience building strategies and a range of new bodies has emerged to provide analysis, training, and resources.</p>
<p>This activity hasn’t been fruitless. We now have a more nuanced understanding of disinformation as a social, psychological, political, and technological phenomenon. Efforts to support public interest journalism and the cultivation of critical thinking through education are also promising. Most notably, major tech companies <a href="https://www.reuters.com/technology/meta-set-up-team-counter-disinformation-ai-abuse-eu-elections-2024-02-26/">no longer pretend to be neutral platforms</a>. </p>
<p>In the meantime, policymakers have rediscovered their duty to <a href="https://commission.europa.eu/strategy-and-policy/priorities-2019-2024/europe-fit-digital-age/digital-services-act_en">regulate technology</a> in the public interest. </p>
<h2>AI and synthetic media</h2>
<p>Regulatory discussions have added urgency now that AI tools to create synthetic media – media partially or fully generated by computers – have gone mainstream. These deepfakes can be used to imitate the voice and appearance of real people. Deepfake media are impressively realistic and do not require much skill or resources. </p>
<p>This is the culmination of the wider digital revolution whereby successive technologies have made high-quality content production accessible to almost anyone. In contrast, regulatory structures and institutional standards for media were mostly designed in an era when only a minority of professionals had access to production.</p>
<p>Political deepfakes can take different forms. The recent Indonesian election saw a <a href="https://edition.cnn.com/2024/02/12/asia/suharto-deepfake-ai-scam-indonesia-election-hnk-intl/index.html">deepfake video “resurrecting” the late President Suharto</a>. This was ostensibly to encourage people to vote, but it was accused of being propaganda because it produced by the political party that he led.</p>
<p>Perhaps a more obvious use of deepfakes is to spread lies about political candidates. For example, <a href="https://ipi.media/slovakia-deepfake-audio-of-dennik-n-journalist-offers-worrying-example-of-ai-abuse/">fake AI-generated audio</a> released days before Slovakia’s parliamentary election in September 2023 attempted to portray the leader of Progressive Slovakia, Michal Šimečka, as having discussed with a journalist how to rig the vote.</p>
<p>Aside from the obvious effort to undermine a political party, it is worth noting how this deepfake, whose origin was unclear, exemplifies wider efforts to scapegoat minorities and demonise mainstream journalism. </p>
<p>Fortunately, in this instance, the audio was not high-quality, which made it quicker and easier for fact checkers to confirm its inauthenticity. However, the integrity of democratic elections cannot rely on the ineptidude of the fakers.</p>
<p>Deepfake audio technology is at a level of <a href="https://www.scientificamerican.com/article/ai-audio-deepfakes-are-quickly-outpacing-detection/">sophistication that makes detection difficult</a>. Deepfake videos still struggle with certain human features, such as the representation of hands, but the technology is still young.</p>
<p>It is also important to note the Slovakian video was released during the final days of the election campaign. This is a prime time to launch disinformation and manipulation attacks because the targets and independent journalists have their hands full and therefore have little time to respond.</p>
<p>If it is also expensive, time-consuming, and difficult to investigate deep fakes, then it’s not clear how electoral commissions, political candidates, the media, or indeed the electorate should respond when potential cases arise. After all, a false accusation from a deepfake can be as troubling as the actual deepfake.</p>
<p>Another way deepfakes could be used to affect elections can be seen in the way they are already widely used to <a href="https://www.euronews.com/next/2023/04/22/a-lifelong-sentence-the-women-trapped-in-a-deepfake-porn-hell">harass and abuse</a> women and girls. This kind of sexual harassment fits an <a href="https://theconversation.com/online-abuse-could-drive-women-out-of-political-life-the-time-to-act-is-now-214301">existing pattern</a> of abuse that limits political participation by women. </p>
<h2>Questioning electoral integrity</h2>
<p>The difficulty is that it’s not yet clear exactly what impact deepfakes could have on elections. It’s very possible we could see other, similar uses of deepfakes in upcoming elections this year. And we could even see deepfakes used in ways not yet conceived of.</p>
<p>But it’s also worth remembering that not all disinformation is high-tech. There are other ways to attack democracy. Rumours and conspiracy theories about the integrity of the electoral process are an insidious trend. <a href="https://www.ft.com/content/1abd7fde-20b4-11e9-a46f-08f9738d6b2b">Electoral fraud is a global concern</a> given that many countries are only democracies in name. </p>
<p>Clearly, social media platforms enable and drive disinformation in many ways, but it is a mistake to assume the problem begins and ends online. One way to think about the challenge of disinformation during upcoming elections is to think about the strength of the systems that are supposed to uphold democracy. </p>
<p>Is there an independent media system capable of providing high quality investigations in the public interest? Are there independent electoral administrators and bodies? Are there independent courts to adjudicate if necessary? </p>
<p>And is there sufficient commitment to democratic values over self interest
amongst politicians and political parties? This year of elections, we may well find out the answer to these questions.</p><img src="https://counter.theconversation.com/content/224786/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Eileen Culloty coordinates the Ireland Hub of the European Digital Media Observatory, which is part-funded by the European Commission to undertake fact-checks, analysis, and media literacy.</span></em></p>As technology has advanced, AI-generated deepfakes have become more convincing.Eileen Culloty, Assistant Professor, School of Communications, Dublin City UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2256232024-03-19T14:07:44Z2024-03-19T14:07:44ZChina: why the country’s economy has hit a wall – and what it plans to do about it<p>China’s <a href="https://www.bbc.co.uk/news/world-asia-68508868">annual parliamentary meetings</a> in Beijing came to a close on March 11. They were conducted under great pressure: a weak economy and high expectations from both the domestic public and international observers as to what the government can do to get the economy out of the woods.</p>
<p>The country’s leaders did not shy away from mentioning all of the economic problems facing China. But they also attempted to boost the morale of the Chinese public by outlining how the country would march into the next chapter of the Chinese story – mainly by striving to become a global leader in technology.</p>
<p>The government used the meetings to <a href="https://npcobserver.com/wp-content/uploads/2024/03/2024-Government-Work-Report_EN.pdf">declare</a> that it was targeting GDP growth of 5% in 2024. This is lower than the <a href="https://www.reuters.com/world/china/chinas-q4-gdp-grows-52-yy-below-market-forecast-2024-01-17/">5.2% growth rate</a> that was achieved in 2023 but higher than the International Monetary Fund’s <a href="https://www.imf.org/en/Countries/CHN">forecast</a> of 4.6%. The Chinese government did not detail how this target will be achieved, but the target itself is indicative of the leadership’s confidence about the future.</p>
<p>Over the past four decades, China’s rapid economic growth has been attributed to market incentives, cheap labour, infrastructure investment, exports and foreign direct investment. But at the time of writing, none of these drivers are working effectively. </p>
<p>Market activities are intertwined with <a href="https://www.piie.com/research/piie-charts/2023/chinas-state-vs-private-company-tracker-which-sector-dominates">greater state intervention</a>. A declining population has weakened the labour supply. And uncertainty surrounding China’s economy and intensified geopolitical tensions have together driven foreign investment <a href="https://asia.nikkei.com/Economy/Foreign-direct-investment-in-China-falls-to-30-yearlow#:%7E:text=But%20FDI%20declined%20for%20the,recorded%20in%20the%20prior%20quarter.">out of China</a>. By January 2024, inward foreign direct investment in China was less than 10% of the US$344 billion (£270 billion) it received in 2021.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/chinas-doom-loop-a-dramatically-smaller-and-older-population-could-create-a-devastating-global-slowdown-221554">China's doom loop: a dramatically smaller (and older) population could create a devastating global slowdown</a>
</strong>
</em>
</p>
<hr>
<h2>Property crisis</h2>
<p>Many of the risks facing China’s economy stem from its ailing real estate sector. For decades, China’s economy was dependent on a <a href="https://www.imf.org/en/News/Articles/2024/02/02/cf-chinas-real-estate-sector-managing-the-medium-term-slowdown#:%7E:text=Real%20estate%20has%20long%20been,the%20buildup%20of%20significant%20risks">booming property market</a> driven by speculative investment returns. However, this growth was largely driven by debt. To maximise their profits, developers even began selling houses before they had been built.</p>
<figure class="align-center ">
<img alt="A view of a room full of people in China sat facing a stage." src="https://images.theconversation.com/files/582141/original/file-20240315-28-bsptnj.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/582141/original/file-20240315-28-bsptnj.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/582141/original/file-20240315-28-bsptnj.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/582141/original/file-20240315-28-bsptnj.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/582141/original/file-20240315-28-bsptnj.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/582141/original/file-20240315-28-bsptnj.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/582141/original/file-20240315-28-bsptnj.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">Delegates attending the closing meeting of the Two Sessions on March 11.</span>
<span class="attribution"><a class="source" href="https://epaimages.com/search.pp">Wu Hao/Shutterstock</a></span>
</figcaption>
</figure>
<p>China’s economy started to slow and, in 2020, Chinese regulators <a href="https://www.bloomberg.com/news/articles/2020-10-08/what-china-s-three-red-lines-mean-for-property-firms-quicktake">cracked down</a> on reckless borrowing. Beijing imposed widespread lending curbs on property developers, meaning they could not borrow more money to pay back their existing debts. </p>
<p>A crisis followed. In early 2024, Evergrande – the world’s most heavily indebted real estate developer – <a href="https://www.reuters.com/business/embattled-china-evergrande-back-court-liquidation-hearing-2024-01-28/">went bust</a>. And other large property developers are in trouble. <a href="https://www.bbc.co.uk/news/business-67142093">Country Garden</a> has defaulted and <a href="https://www.ft.com/content/e1ffbcb4-3222-4a8e-be61-e3a6051567f5">Vanke</a> is struggling to find the new loans it needs to stay alive. </p>
<p>The government confirmed its determination to deflate the property bubble in its annual meeting. It did not highlight how to protect more property developers from defaulting, and only hinted at giving some help to allow developers to complete property projects.</p>
<p>The current weak consumer demand in China’s economy is closely related to the real estate crisis. The value of houses is <a href="https://www.bloomberg.com/news/articles/2024-01-17/china-home-prices-fall-most-since-2015-as-downturn-persists">much lower</a> today than it was two years ago, creating fear about the future value of personal wealth. This has prompted more precautionary saving and less consumption in the face of weak social protection, leading to a general <a href="https://www.theguardian.com/business/2024/feb/08/china-consumer-prices-plunge-at-fastest-rate-for-15-years-as-deflation-fears-deepen">decline in the price</a> of goods and services. </p>
<p>Demand for Chinese goods from abroad has also been <a href="https://www.reuters.com/world/eu-us-pledge-joint-action-over-china-concerns-2023-05-13/">declining</a> due to trade restrictions imposed by the US and the EU, geopolitical concerns and shocks to global supply chains. This explains why throughout its annual meetings the government explicitly emphasised the need to strengthen the self-reliance of the Chinese economy. </p>
<h2>New drivers of growth</h2>
<p>The most eye-catching phrase to come out of the annual meetings was “new quality productive forces”. There are varying interpretations of the term, but they all focus on technology and innovation. </p>
<p>Chinese officials explicitly highlighted the need for China to strive to invent more products related to Artificial Intelligence (AI). The government envisions applications such as AI-powered travel agents and salespeople. </p>
<p>China has, up to this point, been better known for applying AI technologies. Beijing, Shanghai and Shenzhen are all <a href="https://dgap.org/en/research/publications/chinas-smart-cities-and-future-geopolitics">smart cities</a>, where advanced technologies such as AI, cloud computing and big data are used in various areas including transport, urban planning and public security.</p>
<figure class="align-center ">
<img alt="A robot police officer driving down a street in China." src="https://images.theconversation.com/files/582143/original/file-20240315-22-k4w6kp.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/582143/original/file-20240315-22-k4w6kp.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=338&fit=crop&dpr=1 600w, https://images.theconversation.com/files/582143/original/file-20240315-22-k4w6kp.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=338&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/582143/original/file-20240315-22-k4w6kp.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=338&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/582143/original/file-20240315-22-k4w6kp.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=424&fit=crop&dpr=1 754w, https://images.theconversation.com/files/582143/original/file-20240315-22-k4w6kp.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=424&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/582143/original/file-20240315-22-k4w6kp.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=424&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">Shanghai’s first robot police officer patrolling the streets in 2019.</span>
<span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/shanghai-china-dec-20-2019-shanghais-1594426684">atiger/Shutterstock</a></span>
</figcaption>
</figure>
<p>However, transforming China’s economy from one that is driven by investment and fuelled by debt to one that is driven by innovation and technology will bring some fresh challenges. </p>
<p>First, innovation requires incentives and an institutional guarantee to reward risk-taking. Hence, the private sector needs to grow faster. <a href="https://www.piie.com/research/piie-charts/2023/chinas-state-vs-private-company-tracker-which-sector-dominates">Research</a> has found that the share of China’s private sector among the 100 largest listed companies in China dropped to 36.8% at the end of 2023 from 55.4% in mid-2021. </p>
<p>Second, innovation requires more highly skilled human capital. A <a href="https://www.oecd.org/future-of-work/reports-and-data/AI-Employment-brief-2021.pdf">report</a> by the OECD in 2021 concluded that the application of AI technology increases the demand for skilled employees, despite replacing low-skilled labour. This will pose a challenge for China as, up to this point, the country’s growth has been spurred by low-skilled labour. </p>
<p>Third, high-tech industries such as AI and digital services are energy intensive. China has already taken steps to diversify its energy supply, but securing energy supply chains will be essential in the longer term. </p>
<p>Heightened geopolitical tensions and <a href="https://www.bbc.co.uk/news/business-63883047">revamped global supply chains</a> may well reduce exports of energy and other natural resources to China in the future. Many of these resources come from developing economies that have exchanged their resources for China’s infrastructure investment in the past. This is unlikely to be the case in the future.</p><img src="https://counter.theconversation.com/content/225623/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Hong Bo previously received funding from the British Academy. </span></em></p>China is facing many economic obstacles, but Beijing remains optimistic about growth.Hong Bo, Professor of Financial Economics, SOAS, University of LondonLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2202712024-03-19T12:26:28Z2024-03-19T12:26:28ZBuilding fairness into AI is crucial – and hard to get right<figure><img src="https://images.theconversation.com/files/582656/original/file-20240318-18-u3qu8i.jpg?ixlib=rb-1.1.0&rect=0%2C0%2C3478%2C3071&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Are the AIs making decisions about your life fair?</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/illustration/personnel-evaluation-by-artificial-royalty-free-illustration/1733429687?phrase=Artificial+intelligence+hiring">sorbetto/DigitalVision Vectors via Getty Images</a></span></figcaption></figure><p>Artificial intelligence’s capacity to process and analyze vast amounts of data has revolutionized decision-making processes, making operations in <a href="https://doi.org/10.7861%2Ffhj.2021-0095">health care</a>, <a href="https://doi.org/10.1287/mnsc.2016.2644">finance</a>, <a href="https://nij.ojp.gov/topics/articles/using-artificial-intelligence-address-criminal-justice-needs">criminal justice</a> and other sectors of society more efficient and, in many instances, more effective.</p>
<p>With this transformative power, however, comes a significant responsibility: the need to ensure that these technologies are developed and deployed in a manner that is <a href="https://doi.org/10.48550/arXiv.1610.02413">equitable and just</a>. In short, AI needs to be fair. </p>
<p>The pursuit of fairness in AI is not merely an ethical imperative but a requirement in order to foster trust, inclusivity and the <a href="https://www.whitehouse.gov/briefing-room/presidential-actions/2023/10/30/executive-order-on-the-safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence/">responsible advancement of technology</a>. However, ensuring that AI is fair is a major challenge. And on top of that, my research as a computer scientist <a href="https://scholar.google.com/citations?hl=en&user=ASf9Q04AAAAJ&view_op=list_works&sortby=pubdate">who studies AI</a> shows that attempts to ensure fairness in AI can have unintended consequences.</p>
<h2>Why fairness in AI matters</h2>
<p>Fairness in AI has emerged as a <a href="https://www.aies-conference.com/2024/">critical area of focus</a> for researchers, developers and policymakers. It transcends technical achievement, touching on <a href="https://www.whitehouse.gov/briefing-room/presidential-actions/2023/10/30/executive-order-on-the-safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence/">ethical, social and legal dimensions of the technology</a>.</p>
<p>Ethically, fairness is a cornerstone of building trust and acceptance of AI systems. People need to trust that AI decisions that affect their lives – for example, hiring algorithms – are made equitably. Socially, AI systems that embody fairness can help address and mitigate historical biases – for example, those against women and minorities – fostering inclusivity. Legally, embedding fairness in AI systems helps bring those systems into alignment with anti-discrimination laws and regulations around the world.</p>
<p>Unfairness can stem from two primary sources: the input data and the algorithms. Research has shown that input data can <a href="https://dx.doi.org/10.2139/ssrn.2477899">perpetuate bias</a> in various sectors of society. For example, in hiring, algorithms processing data that reflects societal prejudices or lacks diversity can <a href="https://doi.org/10.1145/3351095.3372828">perpetuate “like me” biases</a>. These biases favor candidates who are similar to the decision-makers or those already in an organization. When biased data is then used to train a machine learning algorithm to aid a decision-maker, the algorithm can <a href="http://proceedings.mlr.press/v81/buolamwini18a.html?mod=article_inline&ref=akusion-ci-shi-dai-bizinesumedeia">propagate and even amplify these biases</a>.</p>
<h2>Why fairness in AI is hard</h2>
<p>Fairness is inherently subjective, influenced by cultural, social and personal perspectives. In the context of AI, researchers, developers and policymakers often translate fairness to the idea that algorithms <a href="https://fairmlbook.org/index.html">should not perpetuate or exacerbate</a> existing biases or inequalities.</p>
<p>However, measuring fairness and building it into AI systems is fraught with subjective decisions and technical difficulties. Researchers and policymakers have proposed <a href="https://doi.org/10.48550/arXiv.1610.02413">various definitions of fairness</a>, such as demographic parity, equality of opportunity and individual fairness.</p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/hZG9tyOcyx0?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
<figcaption><span class="caption">Why the concept of algorithmic fairness is so challenging.</span></figcaption>
</figure>
<p>These definitions involve different mathematical formulations and underlying philosophies. They also often conflict, highlighting the <a href="https://doi.org/10.1145/3433949">difficulty of satisfying all fairness criteria</a> simultaneously in practice.</p>
<p>In addition, fairness cannot be distilled into a single metric or guideline. It encompasses a spectrum of considerations including, but not limited to, <a href="https://fairmlbook.org/index.html">equality of opportunity, treatment and impact</a>.</p>
<h2>Unintended effects on fairness</h2>
<p>The multifaceted nature of fairness means that AI systems must be scrutinized at every level of their development cycle, from the initial design and data collection phases to their final deployment and ongoing evaluation. This scrutiny reveals another layer of complexity. AI systems are seldom deployed in isolation. They are used as part of often complex and important decision-making processes, such as making recommendations about hiring or allocating funds and resources, and are subject to many constraints, including <a href="https://dx.doi.org/10.2139/ssrn.4589207">security and privacy</a>.</p>
<p>Research my colleagues and I conducted shows that constraints such as <a href="https://doi.org/10.48550/arXiv.2312.03886">computational resources, hardware types</a> and <a href="https://doi.org/10.24963/ijcai.2021/78">privacy</a> can significantly influence the fairness of AI systems. For instance, the need for computational efficiency can lead to simplifications that inadvertently overlook or misrepresent marginalized groups. </p>
<p>In our study on network pruning – a method to make complex machine learning models smaller and faster – we found that this process <a href="https://doi.org/10.48550/arXiv.2205.13574">can unfairly affect certain groups</a>. This happens because the pruning might not consider how different groups are represented in the data and by the model, leading to biased outcomes.</p>
<p>Similarly, privacy-preserving techniques, while crucial, can obscure the data necessary to identify and mitigate biases or disproportionally affect the outcomes for minorities. For example, when statistical agencies add noise to data to protect privacy, this can <a href="https://doi.org/10.24963/ijcai.2021/78">lead to unfair resource allocation</a> because the added noise affects some groups more than others. This disproportionality can also <a href="https://doi.org/10.24963/ijcai.2022/766">skew decision-making processes</a> that rely on this data, such as resource allocation for public services. </p>
<p>These constraints do not operate in isolation but intersect in ways that compound their impact on fairness. For instance, when privacy measures exacerbate biases in data, it can further amplify existing inequalities. This makes it important to have a comprehensive understanding and approach to both privacy and fairness for AI development.</p>
<h2>The path forward</h2>
<p>Making AI fair is not straightforward, and there are no one-size-fits-all solutions. It requires a process of continuous learning, adaptation and collaboration. Given that bias is pervasive in society, I believe that people working in the AI field should recognize that it’s not possible to achieve perfect fairness and instead strive for continuous improvement. </p>
<p>This challenge requires a commitment to rigorous research, thoughtful policymaking and ethical practice. To make it work, researchers, developers and users of AI will need to ensure that considerations of fairness are woven into all aspects of the AI pipeline, from its conception through data collection and algorithm design to deployment and beyond.</p><img src="https://counter.theconversation.com/content/220271/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Ferdinando Fioretto receives funding from the National Science Foundation, Google, and Amazon. </span></em></p>Bias in AI has been getting a lot of attention lately, but it’s just one aspect of the larger – and thornier – problem of fairness in AI.Ferdinando Fioretto, Assistant Professor of Computer Science, University of VirginiaLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2229872024-03-19T00:12:59Z2024-03-19T00:12:59ZThe ‘digital divide’ is already hurting people’s quality of life. Will AI make it better or worse?<figure><img src="https://images.theconversation.com/files/579788/original/file-20240305-18-nir9gs.jpg?ixlib=rb-1.1.0&rect=17%2C22%2C2775%2C1971&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/road-closed-sign-outback-red-center-1438599635">ChameleonsEye/Shutterstock</a></span></figcaption></figure><p>Today, <a href="https://www.digitalinclusionindex.org.au/wp-content/uploads/2023/07/ADII-2023-Summary_FINAL-Remediated.pdf">almost a quarter of Australians</a> are digitally excluded. This means they miss out on the social, educational and economic benefits <a href="https://ctu.ieee.org/benefits-of-closing-the-global-digital-divide/">online connectivity provides</a>.</p>
<p>In the face of this ongoing “digital divide”, countries are now talking about a future of inclusive artificial intelligence (AI).</p>
<p>However, if we don’t learn from current problems with digital exclusion, it will likely spill over into people’s future experiences with AI. That’s the conclusion from our <a href="https://link.springer.com/article/10.1007/s43681-024-00452-3">new research</a> published in the journal AI and Ethics.</p>
<h2>What is the digital divide?</h2>
<p>The digital divide is a <a href="https://www.sciencedirect.com/science/article/pii/S0040162521007903#bib0030">well-documented social schism</a>. People on the wrong side of it face difficulties when it comes to accessing, affording, or using digital services. These disadvantages significantly reduce their quality of life.</p>
<p><a href="https://www.digitalinclusionindex.org.au/">Decades of research</a> have provided us with a rich understanding of who is most at risk. In Australia, older people, those living in remote areas, people on lower incomes and First Nations peoples are most likely to find themselves digitally excluded.</p>
<p>Zooming out, <a href="https://www.itu.int/itu-d/reports/statistics/facts-figures-2023/">reports</a> show that one-third of the world’s population – representing the poorest countries – remains offline. Globally, the <a href="https://gddindex.com/#:%7E:text=The%20Gender%20Digital%20Divide%20Index%20(GDDI)%20is%20a%20pilot%20benchmarking,gender%20divides%20in%20digital%20development.">digital gender divide</a> also still exists: women, particularly in low and middle-income countries, face substantially more barriers to digital connectivity.</p>
<p>During the COVID pandemic, the impacts of digital inequity became much more obvious. As large swathes of the world’s population had to “shelter in place” – unable to go outside, visit shops, or seek face-to-face contact – anyone without digital access was severely at risk.</p>
<p>Consequences ranged from social isolation to reduced employment opportunities, as well as a lack of access to vital health information. <a href="https://press.un.org/en/2020/sgsm20118.doc.htm">The UN Secretary-General stated in 2020</a> that “the digital divide is now a matter of life and death”. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/579789/original/file-20240305-22-cqzgoc.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="A lonely older woman looking out a window while wearing a medical mask." src="https://images.theconversation.com/files/579789/original/file-20240305-22-cqzgoc.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/579789/original/file-20240305-22-cqzgoc.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/579789/original/file-20240305-22-cqzgoc.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/579789/original/file-20240305-22-cqzgoc.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/579789/original/file-20240305-22-cqzgoc.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/579789/original/file-20240305-22-cqzgoc.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/579789/original/file-20240305-22-cqzgoc.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">People without digital access were severely impacted during the COVID pandemic.</span>
<span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/lonely-senior-woman-surgical-mask-sitting-1688780245">Miriam Doerr Martin Frommherz/Shutterstock</a></span>
</figcaption>
</figure>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/digital-inclusion-and-closing-the-gap-how-first-nations-leadership-is-key-to-getting-remote-communities-online-216085">‘Digital inclusion’ and closing the gap: how First Nations leadership is key to getting remote communities online</a>
</strong>
</em>
</p>
<hr>
<h2>Not just a question of access</h2>
<p>As with most forms of exclusion, the digital divide functions in multiple ways. It was originally defined as a gap between those who have access to computers and the internet and those who do not. But research now shows it’s <a href="https://onlinelibrary.wiley.com/doi/full/10.1111/tesg.12047">not just an issue of access</a>. </p>
<p>Having little or no access leads to reduced familiarity with digital technology, which then erodes confidence, <a href="https://www.cambridge.org/core/books/abs/global-agenda-for-social-justice/tackling-digital-exclusion-counter-social-inequalities-through-digital-inclusion/C9171EE3C4C944FC7712306280EAABDC">fuels disengagement</a>, and ultimately sets in motion <a href="https://www.tandfonline.com/doi/full/10.1080/0144929X.2021.1882577">an intrinsic sense of not being “digitally capable</a>”.</p>
<p>As AI tools increasingly reshape our workplaces, classrooms and everyday lives, there is a risk AI could deepen, rather than narrow, the digital divide.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/artificial-intelligence-holds-great-potential-for-both-students-and-teachers-but-only-if-used-wisely-81024">Artificial intelligence holds great potential for both students and teachers – but only if used wisely</a>
</strong>
</em>
</p>
<hr>
<h2>The role of digital confidence</h2>
<p>To assess the impact of digital exclusion on people’s experiences with AI, in late 2023 we surveyed a representative selection of hundreds of Australian adults. We began by asking them to rate their confidence with digital technology. </p>
<p>We found digital confidence was lower for women, older people, those with reduced salaries, and those with less digital access.</p>
<p>We then asked these same people to comment on their hopes, fears and expectations of AI. Across the board, the data showed that people’s perceptions, attitudes and experiences with AI were linked to how they felt about digital technology in general. </p>
<p>In other words, the more digitally confident people felt, the more positive they were about AI. </p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/giving-ai-direct-control-over-anything-is-a-bad-idea-heres-how-it-could-do-us-real-harm-210168">Giving AI direct control over anything is a bad idea – here's how it could do us real harm</a>
</strong>
</em>
</p>
<hr>
<p>To build truly inclusive AI, these findings are important to consider for several reasons. First, they confirm that digital confidence is not a privilege shared by all. </p>
<p>Second, they show us digital inclusion is about more than just access, or even someone’s digital skills. How confident a person feels in their ability to interact with technology is important too. </p>
<p>Third, they show that if we don’t contend with existing forms of digital exclusion, they are likely to spill over into perceptions, attitudes and experiences with AI. </p>
<p>Currently, <a href="https://www.weforum.org/agenda/2023/09/digital-quality-life-internet-affordability-cybersecurity/">many countries are making headway</a> in their efforts to reduce the digital divide. So we must make sure the rise of AI doesn’t slow these efforts, or worse still, exacerbate the divide.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/577249/original/file-20240222-22-7cjal0.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="A person working on a laptop with the ChatGPT loading screen displayed." src="https://images.theconversation.com/files/577249/original/file-20240222-22-7cjal0.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/577249/original/file-20240222-22-7cjal0.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/577249/original/file-20240222-22-7cjal0.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/577249/original/file-20240222-22-7cjal0.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/577249/original/file-20240222-22-7cjal0.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/577249/original/file-20240222-22-7cjal0.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/577249/original/file-20240222-22-7cjal0.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">AI tools are already transforming lives – but only if you’re on the right side of the ‘digital divide’.</span>
<span class="attribution"><a class="source" href="https://www.pexels.com/photo/a-person-is-using-a-laptop-computer-on-a-table-16094056/">Matheus Bertelli/Pexels</a></span>
</figcaption>
</figure>
<h2>What should we hope for AI?</h2>
<p>While there <a href="https://theconversation.com/forget-dystopian-scenarios-ai-is-pervasive-today-and-the-risks-are-often-hidden-218222">is a slew of associated risks</a>, when deployed responsibly, AI can make significant positive impacts on society. Some of these can directly target issues of inclusivity.</p>
<p>For example, computer vision can <a href="https://www.monash.edu/news/articles/monash-university-and-tennis-australia-serve-up-world-first-accessible-audio-stream-for-fans-with-blindness-or-low-vision">track the trajectory of a tennis ball</a> during a match, making it audible for blind or low-vision spectators.</p>
<p>AI has been used to analyse <a href="https://www.niaa.gov.au/indigenous-affairs/closing-gap/implementation-measures/csiro-indigenous-jobs-map">online job postings</a> to help boost employment outcomes in under-represented populations such as First Nations peoples. And, while they’re still in the early stages of development, <a href="https://www.nature.com/articles/s41746-022-00560-6">AI-powered chatbots</a> could increase accessibility and affordability of medical services. </p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/to-boost-indigenous-employment-we-need-to-map-job-opportunities-to-skills-and-qualifications-our-new-project-does-just-that-212440">To boost Indigenous employment, we need to map job opportunities to skills and qualifications. Our new project does just that</a>
</strong>
</em>
</p>
<hr>
<p>But this responsible AI future can only be delivered if we also address what keeps us digitally divided. To develop and use truly inclusive AI tools, we first have to ensure the feelings of digital exclusion don’t spill over. </p>
<p>This means not only tackling pragmatic issues of access and infrastructure, but also the knock-on effects on people’s levels of engagement, aptitude and confidence with technology.</p><img src="https://counter.theconversation.com/content/222987/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Sarah Bentley works for CSIRO, which receives funding from the Australian Government.</span></em></p><p class="fine-print"><em><span>Claire Naughtin works for CSIRO, which receives funding from the Australian Government.</span></em></p>The benefits of AI are transforming modern life — but disparities in digital confidence are leaving some behind.Sarah Vivienne Bentley, Research Scientist, Responsible Innovation, Data61, CSIROClaire Naughtin, Principal Research Consultant in Strategic Foresight, Data61Licensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2234152024-03-18T12:32:12Z2024-03-18T12:32:12ZAs the US government and record labels go after TikTok, musicians get the squeeze<figure><img src="https://images.theconversation.com/files/582257/original/file-20240315-16-a1ogtt.jpg?ixlib=rb-1.1.0&rect=0%2C24%2C8243%2C5462&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Thomas Raggi of the band Måneskin performs a concert that streamed live on TikTok in 2021.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/news-photo/thomas-raggi-of-the-band-maneskin-performs-at-a-live-news-photo/1233487624?adppopup=true">Fabian Sommer/Picture Alliance via Getty Images</a></span></figcaption></figure><p>For much of the year, TikTok has been on the defensive. </p>
<p>On March 13, 2024, the House of Representatives <a href="https://apnews.com/article/tiktok-ban-house-vote-china-national-security-8fa7258fae1a4902d344c9d978d58a37">voted to approve a bill</a> that would force the short-form video app to be sold off from its Chinese parent company to non-Chinese owners or face a ban in the U.S. The Senate will still have to vote on the legislation, which received broad bipartisan support due to beliefs that TikTok creates risks to national security.</p>
<p>Meanwhile, Universal Music Group, one of the biggest record labels in the world, <a href="https://www.universalmusic.com/an-open-letter-to-the-artist-and-songwriter-community-why-we-must-call-time-out-on-tiktok/">stopped licensing its music to TikTok</a> at the end of January 2024. Since then, songs by Taylor Swift, Billie Eilish and scores of other artists can no longer be used on the platform, while millions of TikTok videos that had incorporated tracks from Universal artists were muted.</p>
<p>Universal Music Group has an estimated <a href="https://seekingalpha.com/article/4580695-universal-music-group-buy-the-leader-of-the-music-industry-ahead-of-earnings">37.5% market share</a> in the music industry, so its songs likely make up a significant portion of the clips used on TikTok prior to the ban.</p>
<p><a href="https://www.nbcnews.com/tech/internet/tiktok-begins-removing-universal-music-publishing-songs-expanding-roya-rcna140713">The record label claims</a> its artists account for a majority of songs on the platform, and therefore, Universal artists should be better compensated and have guardrails against the harmful effects of artificial intelligence. TikTok, in its response, has said that it has come to amicable agreements with other record labels and that Universal is being unreasonable to the detriment of the artists it seeks to protect.</p>
<p>In the end, both companies simply want to have a larger piece of the pie.</p>
<p>But each of their interests, I believe, should be secondary to the creators that sustain them. Over the past two decades, as the internet and streaming have disrupted the music industry, wage gains for music professionals have been far more pronounced at the top of the income ladder. However, most composers and performers have seen their income and employment prospects dwindle.</p>
<p>TikTok has become a beacon in an otherwise dismal digital streaming landscape, and while musicians increasingly need TikTok, TikTok also needs music. </p>
<h2>Gains have gone to the top</h2>
<p><a href="https://www.proquest.com/openview/bc01f8f80efe2e8d006b26520064d146">My research explores the impact of technology</a> on music professionals in the internet era.</p>
<p><a href="https://www.fastcompany.com/91040797/what-the-digital-streaming-revolution-of-the-2000s-can-teach-us-about-the-ai-revolution-today-according-to-a-former-musician">Technology was supposed to democratize the music industry</a>, allowing more artists to more easily gain access to new markets.</p>
<p>Artists no longer needed a record deal to record their music and get it out to the world. They can record music cheaply using their computers, upload it to YouTube, Spotify, BandCamp, SoundCloud, Tidal or any number of platforms for music distribution, then promote their work on social media to build their audience.</p>
<p>But this didn’t lead to more music professionals making a living off their work.</p>
<p>That’s the conclusion I came to by analyzing data from the Bureau of Labor Statistics, which includes two categories of music professionals: performers, who record songs and put on live shows, and composers, which includes musicians who conduct performances or create original works of music but do not necessarily perform that music. A performer would be someone like Dua Lipa, whereas a composer might be someone who is credited for writing a track on Dua Lipa’s album. </p>
<p>From 1999 to 2022, composers saw a strong 85.3% boost in employment, reflecting a gain of 5,380 jobs. This alone suggests that technology has helped music professionals gain employment.</p>
<p>However, when we look at performers – whose employment numbers shrank by 14,690, or 31.6% – it tells a different story. </p>
<p>Put together, the total number of music professionals fell by 9,310 people from 1999 to 2022, reflecting a 17.6% drop. All the free promotion of social media and the lowered barriers to entry that the internet provided were not enough to sustain artists’ livelihoods.</p>
<p><iframe id="ELUqs" class="tc-infographic-datawrapper" src="https://datawrapper.dwcdn.net/ELUqs/2/" height="400px" width="100%" style="border: none" frameborder="0"></iframe></p>
<p>Wages tell a more complex story. </p>
<p>While more people have earned a living from composing music since 1999, their wage gains paled in comparison to that of performers. In short, there are fewer people working as performers now, but those who can cut it are making more money. </p>
<p>This would seem to show that technology has helped most working music professionals.</p>
<p>However, there were outsize gains among the top 10% of music professionals – so the bulk of the rewards from technological advancement went to those at the top. The average wage gain for music professions rises as you climb the income ladder.</p>
<p><iframe id="5G9kt" class="tc-infographic-datawrapper" src="https://datawrapper.dwcdn.net/5G9kt/4/" height="400px" width="100%" style="border: none" frameborder="0"></iframe></p>
<h2>Artists first, or artists last?</h2>
<p>Artists, then, are having an increasingly difficult time making a living, especially independent artists who comprise the lower income brackets. </p>
<p>The promises of technology <a href="https://hbr.org/2024/01/is-genais-impact-on-productivity-overblown">are often overblown</a>; in the case of music, the winners and losers have ended up mirroring <a href="https://www.brookings.edu/articles/rising-inequality-a-major-issue-of-our-time/">broader societal inequalities</a>.</p>
<p>Even as technology hasn’t deliver what it promised to artists, artists are increasingly reliant on technology to make a living. </p>
<p>They’ve increasingly turned to TikTok to do so.</p>
<p>TikTok, with <a href="https://www.statista.com/statistics/272014/global-social-networks-ranked-by-number-of-users/">more than a billion active users</a> worldwide, has revolutionized music promotion and discovery. Unlike traditional social media, TikTok’s unique format, algorithm-driven content discovery and collaborative features supposedly democratize fame. </p>
<p>Lesser-known artists can go viral, shaping the Billboard charts and propelling songs into the mainstream. Lil Nas X <a href="https://www.grammy.com/news/lil-nas-xs-no-1-run-began-tiktok-now-music-industry-taking-notice">rose to fame on TikTok</a> with “Old Town Road” and promptly signed on to Columbia Records. Oliver Anthony, the creator of the populist hit “Rich Men North Of Richmond,” <a href="https://www.billboard.com/music/chart-beat/oliver-anthony-music-rich-men-north-of-richmond-number-one-debut-hot-100-1235396681/">went viral</a> in summer 2023, eventually reaching the No. 1 spot on the Billboard Hot 100.</p>
<p>In this era of virality, TikTok has become an essential promotional tool for musicians and record labels alike, transcending the boundaries of conventional social platforms.</p>
<p>By cutting ties with TikTok, Universal Music Group is not only depriving its artists of these opportunities, but it’s also alienating a large and loyal fan base who use TikTok to interact with their favorite artists and their songs. </p>
<p>TikTok also loses in this situation, since music is such a critical part of its audiovisual experience. <a href="https://www.bloomberg.com/news/articles/2023-03-22/tiktok-lost-customers-when-it-took-away-music-in-australia">In a 2023 test conducted by TikTok</a>, the platform limited the music that some users in Australia could use in posts. For three straight weeks, the number of users, along with the time users spent on the app, declined. </p>
<p>Both parties say they want to protect the artists, <a href="https://newsroom.tiktok.com/en-us/tiktok-statement-in-response-to-universal-music-group">with TikTok arguing</a> that it has reached “artist-first agreements with every other label” and that “Universal’s self-serving actions are not in the best interests of artists, songwriters, and fans.”</p>
<p>TikTok is banking on the perception that platforms provide opportunities for cultural producers by saying that the power of the platform lies in it being “a free promotional and discovery vehicle” for artists. Some members of Congress who opposed the TikTok ban cited the platform’s <a href="https://www.wsj.com/politics/policy/tiktok-bill-ban-house-vote-af4d0800?mod=hp_lead_pos1">utility for maintaining creators’ livelihoods</a>, so this is a common refrain.</p>
<figure class="align-center ">
<img alt="Woman holds sign reading 'I'm 1 of 170 million Americans on TikTok.'" src="https://images.theconversation.com/files/582253/original/file-20240315-28-ytsbjo.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/582253/original/file-20240315-28-ytsbjo.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/582253/original/file-20240315-28-ytsbjo.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/582253/original/file-20240315-28-ytsbjo.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/582253/original/file-20240315-28-ytsbjo.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/582253/original/file-20240315-28-ytsbjo.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/582253/original/file-20240315-28-ytsbjo.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">A protester holds a sign in support of TikTok at a news conference outside the U.S. Capitol on March 12, 2024.</span>
<span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/news-photo/participants-hold-signs-in-support-of-tiktok-at-a-news-news-photo/2079160123?adppopup=true">Anna Moneymaker/Getty Images</a></span>
</figcaption>
</figure>
<p>In response, Universal Music Group has declared that TikTok has an “outdated view” of the modern music business due to the app’s insistence that it provides exposure for artists – and that this exposure is good enough. As my research shows, this free promotion has not grown the ranks of artists who can make a living off music.</p>
<p>TikTok still holds out hope that it can reach “<a href="https://newsroom.tiktok.com/en-us/umpg-update-february-28-2024">an equitable agreement with Universal Music Group</a>,” but the record label hasn’t budged.</p>
<p>The two media companies say they want to protect artists. But I believe the artists are the ones who will end up hurt the most in a divorce.</p>
<p>In other words, TikTok and Universal need to stay together for the kids.</p><img src="https://counter.theconversation.com/content/223415/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Ediz Ozelkan does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.</span></em></p>For some musical artists, TikTok has become a beacon in an otherwise dismal digital streaming landscape.Ediz Ozelkan, Lecturer of Media Studies, University of Colorado BoulderLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2200362024-03-18T12:31:28Z2024-03-18T12:31:28ZAI vs. elections: 4 essential reads about the threat of high-tech deception in politics<figure><img src="https://images.theconversation.com/files/582204/original/file-20240315-28-p5czjg.jpg?ixlib=rb-1.1.0&rect=0%2C0%2C4977%2C6250&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Like it or not, AI is already playing a role in the 2024 presidential election.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/android-celebrating-4th-july-royalty-free-image/499467267?phrase=Robot+Uncle+Sam">kirstypargeter/iStock via Getty Images</a></span></figcaption></figure><p>It’s official. Joe Biden and Donald Trump have <a href="https://www.washingtonpost.com/politics/2024/03/13/few-voters-decide-trump-biden-nominations/">secured the necessary delegates</a> to be their parties’ nominees for president in the 2024 election. Barring unforeseen events, the two will be formally nominated at the party conventions this summer and face off at the ballot box on Nov. 5. </p>
<p>It’s a safe bet that, as <a href="https://theconversation.com/how-tech-firms-have-tried-to-stop-disinformation-and-voter-intimidation-and-come-up-short-148771">in recent elections</a>, this one will play out largely online and feature a potent blend of news and disinformation delivered over social media. New this year are powerful generative artificial intelligence tools such as <a href="https://openai.com/chatgpt">ChatGPT</a> and <a href="https://openai.com/sora">Sora</a> that make it easier to “<a href="https://ssrn.com/abstract=4040800">flood the zone</a>” with propaganda and disinformation and produce convincing deepfakes: words coming from the mouths of politicians that they did not actually say and events replaying before our eyes that did not actually happen.</p>
<p>The result is an increased likelihood of voters being deceived and, perhaps as worrisome, a growing sense that <a href="https://www.researchgate.net/publication/378236203_Profiling_the_Dynamics_of_Trust_Distrust_in_Social_Media_A_Survey_Study">you can’t trust anything you see online</a>. Trump is already taking advantage of the so-called <a href="https://doi.org/10.1017/S0003055423001454">liar’s dividend</a>, the opportunity to discount your actual words and deeds as deepfakes. Trump implied on his Truth Social platform on March 12, 2024, that real videos of him shown by Democratic House members were <a href="https://www.washingtonpost.com/politics/2024/03/13/trump-video-ai-truth-social/">produced or altered using artificial intelligence</a>.</p>
<p>The Conversation has been covering the latest developments in artificial intelligence that have the potential to undermine democracy. The following is a roundup of some of those articles from our archive. </p>
<h2>1. Fake events</h2>
<p>The ability to use AI to make convincing fakes is particularly troublesome for producing false evidence of events that never happened. Rochester Institute of Technology computer security researcher <a href="https://scholar.google.com/citations?user=UxGWcUYAAAAJ&hl=en">Christopher Schwartz</a> has dubbed these <a href="https://theconversation.com/events-that-never-happened-could-influence-the-2024-presidential-election-a-cybersecurity-researcher-explains-situation-deepfakes-206034">situation deepfakes</a>.</p>
<p>“The basic idea and technology of a situation deepfake are the same as with any other deepfake, but with a bolder ambition: to manipulate a real event or invent one from thin air,” he wrote.</p>
<p>Situation deepfakes could be used to boost or undermine a candidate or suppress voter turnout. If you encounter reports on social media of events that are surprising or extraordinary, try to learn more about them from reliable sources, such as fact-checked news reports, peer-reviewed academic articles or interviews with credentialed experts, Schwartz said. Also, recognize that deepfakes can take advantage of what you are inclined to believe.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/events-that-never-happened-could-influence-the-2024-presidential-election-a-cybersecurity-researcher-explains-situation-deepfakes-206034">Events that never happened could influence the 2024 presidential election – a cybersecurity researcher explains situation deepfakes</a>
</strong>
</em>
</p>
<hr>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/S4gd-EpBlS0?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
<figcaption><span class="caption">How AI puts disinformation on steroids.</span></figcaption>
</figure>
<h2>2. Russia, China and Iran take aim</h2>
<p>From the question of what AI-generated disinformation can do follows the question of who has been wielding it. Today’s AI tools put the capacity to produce disinformation in reach for most people, but of particular concern are <a href="https://theconversation.com/ai-disinformation-is-a-threat-to-elections-learning-to-spot-russian-chinese-and-iranian-meddling-in-other-countries-can-help-the-us-prepare-for-2024-214358">nations that are adversaries</a> of the United States and other democracies. In particular, Russia, China and Iran have extensive experience with disinformation campaigns and technology.</p>
<p>“There’s a lot more to running a disinformation campaign than generating content,” wrote security expert and Harvard Kennedy School lecturer <a href="https://www.schneier.com/">Bruce Schneier</a>. “The hard part is distribution. A propagandist needs a series of fake accounts on which to post, and others to boost it into the mainstream where it can go viral.”</p>
<p>Russia and China have a history of testing disinformation campaigns on smaller countries, according to Schneier. “Countering new disinformation campaigns requires being able to recognize them, and recognizing them requires looking for and cataloging them now,” he wrote.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/ai-disinformation-is-a-threat-to-elections-learning-to-spot-russian-chinese-and-iranian-meddling-in-other-countries-can-help-the-us-prepare-for-2024-214358">AI disinformation is a threat to elections − learning to spot Russian, Chinese and Iranian meddling in other countries can help the US prepare for 2024</a>
</strong>
</em>
</p>
<hr>
<h2>3. Healthy skepticism</h2>
<p>But it doesn’t require the resources of shadowy intelligence services in powerful nations to make headlines, as the New Hampshire <a href="https://apnews.com/article/ai-robocall-biden-new-hampshire-primary-2024-f94aa2d7f835ccc3cc254a90cd481a99">fake Biden robocall</a> produced and disseminated by two individuals and aimed at dissuading some voters illustrates. That episode prompted the Federal Communications Commission to <a href="https://theconversation.com/fcc-bans-robocalls-using-deepfake-voice-clones-but-ai-generated-disinformation-still-looms-over-elections-223160">ban robocalls that use voices generated</a> by artificial intelligence. </p>
<p>AI-powered disinformation campaigns are difficult to counter because they can be delivered over different channels, including robocalls, social media, email, text message and websites, which complicates the digital forensics of tracking down the sources of the disinformation, wrote <a href="https://scholar.google.com/citations?hl=en&user=yu4Ew7gAAAAJ&view_op=list_works&sortby=pubdate">Joan Donovan</a>, a media and disinformation scholar at Boston University.</p>
<p>“In many ways, AI-enhanced disinformation such as the New Hampshire robocall poses the same problems as every other form of disinformation,” Donovan wrote. “People who use AI to disrupt elections are likely to do what they can to hide their tracks, which is why it’s necessary for the public to remain skeptical about claims that do not come from verified sources, such as local TV news or social media accounts of reputable news organizations.”</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/fcc-bans-robocalls-using-deepfake-voice-clones-but-ai-generated-disinformation-still-looms-over-elections-223160">FCC bans robocalls using deepfake voice clones − but AI-generated disinformation still looms over elections</a>
</strong>
</em>
</p>
<hr>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/L0X3W1utdRQ?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
<figcaption><span class="caption">How to spot AI-generated images.</span></figcaption>
</figure>
<h2>4. A new kind of political machine</h2>
<p>AI-powered disinformation campaigns are also difficult to counter because they can include bots – automated social media accounts that pose as real people – and can include <a href="https://theconversation.com/how-ai-could-take-over-elections-and-undermine-democracy-206051">online interactions tailored to individuals</a>, potentially over the course of an election and potentially with millions of people.</p>
<p>Harvard political scientist <a href="https://scholar.google.com/citations?user=3Bl9cn8AAAAJ&hl=en">Archon Fung</a> and legal scholar <a href="https://scholar.google.com/citations?user=LxG5YWcAAAAJ&hl=en">Lawrence Lessig</a> described these capabilities and laid out a hypothetical scenario of national political campaigns wielding these powerful tools.</p>
<p>Attempts to block these machines could run afoul of the free speech protections of the First Amendment, according to Fung and Lessig. “One constitutionally safer, if smaller, step, already adopted in part by European internet regulators and in California, is to prohibit bots from passing themselves off as people,” they wrote. “For example, regulation might require that campaign messages come with disclaimers when the content they contain is generated by machines rather than humans.”</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/how-ai-could-take-over-elections-and-undermine-democracy-206051">How AI could take over elections – and undermine democracy</a>
</strong>
</em>
</p>
<hr>
<p><em>This story is a roundup of articles from The Conversation’s archives.</em></p>
<hr>
<p><em><strong><a href="https://theconversation.com/us/topics/election-2024-disinformation-151606">This article is part of Disinformation 2024:</a></strong> a series examining the science, technology and politics of deception in elections.</em></p>
<p><em>You may also be interested in:</em></p>
<p><a href="https://theconversation.com/disinformation-is-rampant-on-social-media-a-social-psychologist-explains-the-tactics-used-against-you-216598">Disinformation is rampant on social media – a social psychologist explains the tactics used against you</a></p>
<p><a href="https://theconversation.com/misinformation-disinformation-and-hoaxes-whats-the-difference-158491">Misinformation, disinformation and hoaxes: What’s the difference?</a></p>
<p><a href="https://theconversation.com/disinformation-campaigns-are-murky-blends-of-truth-lies-and-sincere-beliefs-lessons-from-the-pandemic-140677">Disinformation campaigns are murky blends of truth, lies and sincere beliefs – lessons from the pandemic</a></p>
<hr><img src="https://counter.theconversation.com/content/220036/count.gif" alt="The Conversation" width="1" height="1" />
Using disinformation to sway elections is nothing new. Powerful new AI tools, however, threaten to give the deceptions unprecedented reach.Eric Smalley, Science + Technology EditorLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2255092024-03-18T12:28:06Z2024-03-18T12:28:06ZAmid growth in AI writing tools, this course teaches future lawyers and other professionals to become better editors<figure><img src="https://images.theconversation.com/files/581594/original/file-20240313-18-ljzu6.jpg?ixlib=rb-1.1.0&rect=20%2C93%2C6852%2C4260&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Even the best paragraphs may have room for improvement.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/cyber-law-or-internet-law-concept-with-ai-robot-royalty-free-image/1350320510?phrase=law+students+writing+ai+&adppopup=true">PhonlamaiPhoto via Getty Images</a></span></figcaption></figure><figure class="align-right ">
<img alt="Text saying: Uncommon Courses, from The Conversation" src="https://images.theconversation.com/files/499014/original/file-20221205-17-kcwec8.png?ixlib=rb-1.1.0&q=45&auto=format&w=237&fit=clip" srcset="https://images.theconversation.com/files/499014/original/file-20221205-17-kcwec8.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=375&fit=crop&dpr=1 600w, https://images.theconversation.com/files/499014/original/file-20221205-17-kcwec8.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=375&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/499014/original/file-20221205-17-kcwec8.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=375&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/499014/original/file-20221205-17-kcwec8.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=471&fit=crop&dpr=1 754w, https://images.theconversation.com/files/499014/original/file-20221205-17-kcwec8.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=471&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/499014/original/file-20221205-17-kcwec8.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=471&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption"></span>
</figcaption>
</figure>
<p><em><a href="https://theconversation.com/topics/uncommon-courses-130908">Uncommon Courses</a> is an occasional series from The Conversation U.S. highlighting unconventional approaches to teaching.</em> </p>
<h2>Title of Course:</h2>
<p>“Editing and Advocacy”</p>
<h2>What prompted the idea for the course?</h2>
<p>In part, I wanted to improve the career prospects of the law students, business students and other aspiring professionals I teach. People who can consistently improve the sentences and paragraphs that come across their desk each day have the opportunity to improve the way ideas and messages are communicated. Who wouldn’t want to add someone like that to their company, government agency or nonprofit organization?</p>
<p>Mostly, though, I designed the course so that my students can experience the empowering magic that comes with being able to take a string of words — whether drafted by themselves or somebody else — and transform them into a revised version that is undeniably better than the original. </p>
<h2>What does the course explore?</h2>
<p>Students edit emails. They edit contracts. They edit memos, articles, speeches, proposals, text messages, blog posts — pretty much anything that lawyers and other professionals compose. Sometimes they edit alone. Other times they edit as part of a team. But the goal is always the same: learn and practice a skill that is fundamental to becoming an excellent advocate.</p>
<h2>Why is this course relevant now?</h2>
<p>I first starting teaching “Editing and Advocacy” a few years before the launch of ChatGPT and other generative AI tools. But now that those tools have significantly reduced the cost of producing drafts, the course’s focus on revising drafts — for accuracy, for clarity, for persuasive power — has taken on a newfound relevance.</p>
<p>For instance, when asked how AI might affect what he and other members of the knowledge economy do, tech journalist <a href="https://www.theringer.com/2023/3/21/23649894/the-ai-revolution-could-be%2520-bigger-and-weirder-than-we-can-imagine.">Charlie Warzel suggested</a> that “the greatest skill that we can all have now is to be ‘editors.’” We may, he noted, start to spend an increasing amount of time correcting and refining AI-produced material.</p>
<p>Ilona Logvinova, associate general counsel and head of innovation for the legal department at consulting giant McKinsey, <a href="https://wsjcustomevents.com/lexisnexis2024">made a similar point</a>, telling attendees at a recent conference on the use of AI in law: “I really believe that we’re at a moment where we, as lawyers, can transition from being ‘drafters’ to being ‘editors.’”</p>
<h2>What’s a critical lesson from the course?</h2>
<p>One of the most critical lessons is something I put front and center on the syllabus: “Good editors don’t just see the sentence that was written. They see the sentence that might have been written. They know how to spot words that shouldn’t be included and summon up ones that haven’t yet appeared. Their value comes not just from preventing mistakes but also from discovering new ways to improve a piece’s style, structure, and overall impact.”</p>
<p>The current generation of AI tools is really good at proofreading. But so far, I haven’t encountered any large language model that has the vision, empathy and deep understanding of both context and nuance — not to mention of personal voice —required of a truly exceptional editor.</p>
<figure class="align-center ">
<img alt="A woman peers into a book while seated at her desk, which has both a desktop and a laptop computer." src="https://images.theconversation.com/files/582048/original/file-20240314-21-w762u1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/582048/original/file-20240314-21-w762u1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=338&fit=crop&dpr=1 600w, https://images.theconversation.com/files/582048/original/file-20240314-21-w762u1.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=338&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/582048/original/file-20240314-21-w762u1.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=338&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/582048/original/file-20240314-21-w762u1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=424&fit=crop&dpr=1 754w, https://images.theconversation.com/files/582048/original/file-20240314-21-w762u1.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=424&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/582048/original/file-20240314-21-w762u1.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=424&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">Experts say editing will take on greater importance in the age of artificial intelligence.</span>
<span class="attribution"><span class="source">Laurence Dutton via Getty Images</span></span>
</figcaption>
</figure>
<p>That doesn’t mean that a technology with those capacities won’t eventually develop, nor that the technology we already have can’t provide enormously useful editing assistance. In fact, more and more of my assignments in “Editing and Advocacy” give students a chance to play around with ChatGPT-like tools. I have also created an entirely separate course called “Digital Lawyering: Advocacy in the Age of AI” that explores the possibilities – and pitfalls – of using artificial intelligence as a kind of co-counsel.</p>
<p>But as I often remind students in both classes, editing is as much about imagination, emotional intelligence and restraint as it is about syntax, semicolons and subject-and-verb agreement. A good way to become better at it is to cultivate the parts of you that are most human. </p>
<h2>What materials does the course feature?</h2>
<p>Hoping to save my students some money — and wanting to make the materials of the course easily available online — I worked with the <a href="https://www.publishing.umich.edu/our-mission">publishing team</a> at the University of Michigan to create a set of open-access books that anyone with an internet connection can read for free. These include “<a href="https://www.fulcrum.org/concern/monographs/hq37vr12w">Editing and Advocacy</a>,” “<a href="https://www.fulcrum.org/concern/monographs/dv13zw31v">Notes on Nuance</a>,” “<a href="https://www.fulcrum.org/concern/monographs/f1881p37d">Punctuation and Persuasion</a>” and “<a href="https://www.fulcrum.org/concern/monographs/8623j145m">Feedback Loops: How to Give and Receive High-Quality Feedback</a>.”</p>
<h2>What will the course prepare students to do?</h2>
<p>Editing involves reliably making informed, value-creating decisions. You need to know what to add. You need to know what to delete. You need to know what to separate, combine and rearrange. Students in the course study, evaluate and regularly participate in those types of decisions. In the process, they develop an extremely important and highly transferable skill: good judgment.</p><img src="https://counter.theconversation.com/content/225509/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Patrick Barry does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.</span></em></p>Learning how to produce polished prose can greatly enhance your value on the job.Patrick Barry, Clinical Assistant Professor and Director of Digital Academic Initiatives (University of Michigan Law School) | Visiting Lecturer (University of Chicago Law School) | Visiting Lecturer (UCLA School of Law), University of MichiganLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2255552024-03-17T19:01:36Z2024-03-17T19:01:36ZSomething felt ‘off’ – how AI messed with our human research, and what we learned<p>All levels of research are being changed by the rise of artificial intelligence (AI). Don’t have time to read that journal article? AI-powered tools such as <a href="https://www.tldrthis.com/">TLDRthis</a> will summarise it for you. </p>
<p>Struggling to find relevant sources for your review? <a href="https://inciteful.xyz/">Inciteful</a> will list suitable articles with just the click of a button. Are your human research participants too expensive or complicated to manage? Not a problem – try <a href="https://www.syntheticusers.com/">synthetic participants</a> instead. </p>
<p>Each of these tools suggests AI could be superior to humans in outlining and explaining concepts or ideas. But can humans be replaced when it comes to qualitative research?</p>
<p>This is something we recently had to grapple with while carrying out unrelated research into <a href="https://compass.onlinelibrary.wiley.com/doi/abs/10.1111/spc3.12643">mobile dating during the COVID-19 pandemic</a>. And what we found should temper enthusiasm for artificial responses over the words of human participants.</p>
<p><div data-react-class="Tweet" data-react-props="{"tweetId":"1768426948638675367"}"></div></p>
<h2>Encountering AI in our research</h2>
<p>Our research is looking at how people might navigate mobile dating during the pandemic in Aotearoa New Zealand. Our aim was to explore broader social responses to mobile dating as the pandemic progressed and as public health mandates changed over time.</p>
<p>As part of this ongoing research, we prompt participants to develop stories in response to hypothetical scenarios. </p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/what-happens-when-we-outsource-boring-but-important-work-to-ai-research-shows-we-forget-how-to-do-it-ourselves-223981">What happens when we outsource boring but important work to AI? Research shows we forget how to do it ourselves</a>
</strong>
</em>
</p>
<hr>
<p>In 2021 and 2022 we received a wide range of intriguing and quirky responses from 110 New Zealanders recruited through Facebook. Each participant received a gift voucher for their time.</p>
<p>Participants described characters navigating the challenges of “Zoom dates” and clashing over vaccination statuses or wearing masks. Others wrote passionate love stories with eyebrow-raising details. Some even broke the fourth wall and wrote directly to us, complaining about the mandatory word length of their stories or the quality of our prompts. </p>
<p>These responses captured the highs and lows of online dating, the boredom and loneliness of lockdown, and the thrills and despair of finding love during the time of COVID-19. </p>
<p>But, perhaps most of all, these responses reminded us of the idiosyncratic and irreverent aspects of human participation in research – the unexpected directions participants go in, or even the unsolicited feedback you can receive when doing research. </p>
<p>But in the latest round of our study in late 2023, something had clearly changed across the 60 stories we received.</p>
<p>This time many of the stories felt “off”. Word choices were quite stilted or overly formal. And each story was quite moralistic in terms of what one “should” do in a situation. </p>
<p>Using AI detection tools, such as ZeroGPT, we concluded participants – or even bots – were using AI to generate story answers for them, possibly to receive the gift voucher for minimal effort.</p>
<p>Contrary to claims that AI can sufficiently replicate human participants in research, we found AI-generated stories to be woeful. </p>
<p>We were reminded that an essential ingredient of any social research is for the data to be based on lived experience. </p>
<h2>Is AI the problem?</h2>
<p>Perhap the biggest threat to human research is not AI, but rather the philosophy that underscores it. </p>
<p>It is worth noting the majority of claims about AI’s capabilities to replace humans come from computer scientists or quantitative social scientists. In these types of studies, human reasoning or behaviour is often measured through scorecards or yes/no statements. </p>
<p>This approach necessarily fits human experience into a framework that can be more easily analysed through computational or artificial interpretation. </p>
<p>In contrast, we are qualitative researchers who are interested in the messy, emotional, lived experience of people’s perspectives on dating. We were drawn to the thrills and disappointments participants originally pointed to with online dating, the frustrations and challenges of trying to use dating apps, as well as the opportunities they might create for intimacy during a time of lockdowns and evolving health mandates. </p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/ai-is-in-danger-of-becoming-too-male-new-research-121229">AI is in danger of becoming too male – new research</a>
</strong>
</em>
</p>
<hr>
<p>In general, we found AI poorly simulated these experiences. </p>
<p>Some might accept generative AI is here to stay, or that AI should be viewed as offering various tools to researchers. Other researchers might retreat to forms of data collection, such as surveys, that might minimise the interference of unwanted AI participation. </p>
<p>But, based on <a href="https://www.tandfonline.com/doi/full/10.1080/14780887.2024.2311427">our recent research experience</a>, we believe theoretically-driven, qualitative social research is best equipped to detect and protect against AI interference. </p>
<p>There are additional implications for research. The threat of AI as an unwanted participant means researchers will have to work longer or harder to spot imposter participants. </p>
<p>Academic institutions need to start developing policies and practices to reduce the burden on individual researchers trying to carry out research in the changing AI environment. </p>
<p>Regardless of researchers’ theoretical orientation, how we work to limit the involvement of AI is a question for anyone interested in understanding human perspectives or experiences. If anything, the limitations of AI reemphasise the importance of being human in social research.</p><img src="https://counter.theconversation.com/content/225555/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Alexandra Gibson receives funding from Te Apārangi - Royal Society of New Zealand.</span></em></p><p class="fine-print"><em><span>Alex Beattie does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.</span></em></p>Responses to our qualitative survey suggested artificial intelligence was at play. The results were woeful, and researchers will need to work harder to prevent contaminated outcomes.Alexandra Gibson, Senior Lecturer in Health Psychology, Te Herenga Waka — Victoria University of WellingtonAlex Beattie, Research Fellow, School of Health, Te Herenga Waka — Victoria University of WellingtonLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2250362024-03-14T17:19:11Z2024-03-14T17:19:11ZIs it ethical to watch AI pornography?<figure><img src="https://images.theconversation.com/files/580230/original/file-20240306-30-un3efx.jpg?ixlib=rb-1.1.0&rect=79%2C88%2C5811%2C3850&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/mobile-addict-man-using-smartphone-browsing-2391001945">Lysenko Andrii/Shutterstock</a></span></figcaption></figure><p>If you’re in your 20s and 30s, you probably watch pornography. Millennials and gen Z are <a href="https://www.lelo.com/blog/wp-content/uploads/2023/11/UK-Sex-Census-2023.pdf">watching more</a> pornography than any other age group and are also <a href="https://www.lelo.com/blog/wp-content/uploads/2023/11/UK-Sex-Census-2023.pdf">more likely</a> than any other demographic to experiment with AI pornography. </p>
<p>As technology advances, AI-generated tools and techniques are becoming increasingly sophisticated and accessible. This can lead to unethical content, including deepfakes – videos in which a person’s face is replaced with someone
else’s likeness, without their consent. Social media platform X (formerly Twitter) recently faced a scandal when it became awash with <a href="https://www.theguardian.com/technology/2024/jan/31/inside-the-taylor-swift-deepfake-scandal-its-men-telling-a-powerful-woman-to-get-back-in-her-box">deepfakes of Taylor Swift</a>.</p>
<p>But what about other kinds of AI pornographic content? How can consuming it affect you, and how can you make sure that you’re consuming it ethically? I’m a sex and relationship therapist, so I’m interested in helping clients with various sexual issues, including porn consumption problems. I am also curious about the ways AI could be used positively to create pornography that is not only ethical, but educational and sexy at the same time.</p>
<hr>
<figure class="align-right ">
<img alt="Quarter life, a series by The Conversation" src="https://images.theconversation.com/files/451343/original/file-20220310-13-1bj6csd.png?ixlib=rb-1.1.0&q=45&auto=format&w=237&fit=clip" srcset="https://images.theconversation.com/files/451343/original/file-20220310-13-1bj6csd.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=600&fit=crop&dpr=1 600w, https://images.theconversation.com/files/451343/original/file-20220310-13-1bj6csd.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=600&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/451343/original/file-20220310-13-1bj6csd.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=600&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/451343/original/file-20220310-13-1bj6csd.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=754&fit=crop&dpr=1 754w, https://images.theconversation.com/files/451343/original/file-20220310-13-1bj6csd.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=754&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/451343/original/file-20220310-13-1bj6csd.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=754&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption"></span>
</figcaption>
</figure>
<p><em><strong><a href="https://theconversation.com/uk/topics/quarter-life-117947?utm_source=TCUK&utm_medium=linkback&utm_campaign=UK+YP2022&utm_content=InArticleTop">This article is part of Quarter Life</a></strong>, a series about issues affecting those of us in our 20s and 30s. From the challenges of beginning a career and taking care of our mental health, to the excitement of starting a family, adopting a pet or just making friends as an adult. The articles in this series explore the questions and bring answers as we navigate this turbulent period of life.</em></p>
<p><em>You may be interested in:</em></p>
<p><em><a href="https://theconversation.com/what-you-should-know-about-coming-out-as-lgbtq-in-your-20s-and-30s-223910?utm_source=TCUK&utm_medium=linkback&utm_campaign=UK+YP2022&utm_content=InArticleTop">What you should know about coming out as LGBTQ+ in your 20s and 30s</a></em></p>
<p><em><a href="https://theconversation.com/the-problems-with-dating-apps-and-how-they-could-be-fixed-two-relationship-experts-discuss-218401?utm_source=TCUK&utm_medium=linkback&utm_campaign=UK+YP2022&utm_content=InArticleTop">The problems with dating apps and how they could be fixed – two relationship experts discuss</a></em></p>
<p><em><a href="https://theconversation.com/why-you-might-start-to-hate-the-influencers-you-once-loved-222659?utm_source=TCUK&utm_medium=linkback&utm_campaign=UK+YP2022&utm_content=InArticleTop">Why you might start to hate the influencers you once loved</a></em></p>
<hr>
<h2>The impact of watching AI porn</h2>
<p>While it’s perfectly normal to be curious about sex, watching a lot of pornography can affect your <a href="https://www.mdpi.com/2077-0383/8/7/914">sexual satisfaction</a> – and AI porn is no different. You might, for example, start comparing your partner to the hyper-realistic, but impossibly perfect, digitally generated actors of AI porn. </p>
<p>Already, <a href="https://akjournals.com/view/journals/2006/5/2/article-p179.xml">research</a> suggests that men who frequently watch online porn may experience erectile dysfunction. This could be due to the idealised unrealistic portrayals in pornography compared to real-life sexual encounters. AI pornography would likely only exacerbate this, with AI porn avatars able to participate in sex acts that wouldn’t be possible, or as accessible, for real people. </p>
<figure class="align-center ">
<img alt="Young asian woman lying in bed lit by glow of her phone" src="https://images.theconversation.com/files/581005/original/file-20240311-20-jwbnkx.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/581005/original/file-20240311-20-jwbnkx.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/581005/original/file-20240311-20-jwbnkx.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/581005/original/file-20240311-20-jwbnkx.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/581005/original/file-20240311-20-jwbnkx.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/581005/original/file-20240311-20-jwbnkx.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/581005/original/file-20240311-20-jwbnkx.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">Research has found that watching porn can help some women to overcome shame.</span>
<span class="attribution"><a class="source" href="https://www.shutterstock.com/search/woman-phone-bed">TORWAISTUDIO/Shutterstock</a></span>
</figcaption>
</figure>
<p>Among women who watch porn, opinions vary. Some women have <a href="https://www.tandfonline.com/doi/full/10.1080/0092623X.2021.1885532">noted</a> positive changes, including a reduction in the shame associated with sexual pleasure. But others have expressed reservations about the <a href="https://www.researchgate.net/publication/328265584_Is_Women%27s_Problematic_Pornography_Viewing_Related_to_Body_Image_or_Relationship_Satisfaction">beauty standards</a> in pornography, finding them unattainable. </p>
<p>This stance is very much shared by <a href="https://www.sciencedirect.com/science/article/pii/S0277539500000777?via=ihub#BIB30">anti-porn campaigners</a>, who claim that porn degrades and objectifies women. They believe it feeds into ideas of male supremacy, potentially leading to <a href="https://www.researchgate.net/publication/342907785_A_Descriptive_Analysis_of_the_Types_Targets_and_Relative_Frequency_of_Aggression_in_Mainstream_Pornography/link/5f21c47b299bf134049257f8/download">violence against women</a>. </p>
<h2>Regulating AI pornography</h2>
<p>The roles women are given within mainstream pornography often portray a different power imbalance than, for example, gay male porn. For this article, I spoke to porn actor John Thomas. He argued that gay male porn was somewhat more ethical than mainstream straight porn which is [made for men to consume]. “Both roles in a gay scene might be appreciated by the viewer, rather than the pure objectification of the woman in a straight scene.”</p>
<p>One of the many concerns with unregulated AI-generated pornography is that it can distort a viewer’s sense of reality, leading to misinformation, unrealistic expectations around sex and potential harm. However, since the moral landscape surrounding ethical AI porn is a grey area, we are pushed into uncharted territory. As new technology emerges, new challenges arise.</p>
<p>To ensure <a href="https://www.redalyc.org/pdf/3230/323047487009.pdf">responsible</a> innovation within the adult entertainment industry, it’s essential to be aware of AI’s ongoing <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7296407/">integration</a> into our daily lives. For example, risks could be mitigated by training AI systems to recognise deepfakes, violence or child pornography. </p>
<p>For actors in the adult entertainment industry working pre-AI, consent has always been key. I asked John Thomas about best practice in the industry: </p>
<blockquote>
<p>When working for a porn studio [as a freelance worker], I sign a contract which typically includes clauses relating to rights to my image – usually I am signing to give the studio the right to use, and alter, my image [from the photos or video created] and distribute it. AI is not specified in any contract I’ve signed.</p>
</blockquote>
<p>But since AI porn is expected to become more mainstream, the topic of consent becomes more ambiguous. As John Thomas adds: “I think one could interpret the contract to include AI … the contracts are written in such an expansive way that, having signed away the rights to your image, and consent to your image being altered, it could [hypothetically be used in AI].”</p>
<h2>How to be an ethical porn consumer</h2>
<p>Just as there are fair trade brands known for their ethical practices in producing coffee and clothing, there should be a safe space for consumers to explore their sexuality and fantasies. </p>
<p>As a porn viewer, you can be more ethical in your consumption by becoming <a href="https://www.tandfonline.com/doi/abs/10.1080/00224499.2018.1556238">porn literate</a>, improving your understanding of realistic sexual expectations, gender identities, sexual orientations, relationship styles, kinks and ethical BDSM practices. </p>
<figure class="align-center ">
<img alt="A gay couple in bed together, looking happily at a phone" src="https://images.theconversation.com/files/581293/original/file-20240312-16-wqqfqo.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/581293/original/file-20240312-16-wqqfqo.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/581293/original/file-20240312-16-wqqfqo.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/581293/original/file-20240312-16-wqqfqo.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/581293/original/file-20240312-16-wqqfqo.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/581293/original/file-20240312-16-wqqfqo.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/581293/original/file-20240312-16-wqqfqo.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">It’s good to talk about your porn preferences with your partner.</span>
<span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/affectionate-gay-couple-watching-content-online-2233406561">Lomb/Shutterstock</a></span>
</figcaption>
</figure>
<p>And if you decide you want to watch AI porn and want to <a href="https://journals.sagepub.com/doi/full/10.1177/1363460720936475">minimise</a> the risk of consuming unethical content, here are some tips to help enhance your porn literacy skills: </p>
<p>• consider joining online communities where discussions about <a href="https://journals.sagepub.com/doi/10.1177/1363460720936475?icid=int.sj-full-text.citing-articles.43">“feminist porn”</a> and sexualised content are open and encouraged</p>
<p>• if you are a fan of a particular porn actor, consider following them on social media. This will provide you with some insights into their performance activities and their preferred ways for you to access their content </p>
<p>• when coming across porn sites, take a moment to assess if they are recognised for ethical production practices. Some <a href="https://ses.library.usyd.edu.au/handle/2123/30093">established sites</a> are known for their commitment to ethical pornography. Typically, the ethical emphasis will revolve around aspects such as production standards, consent, representation of diverse body types, genders and races, portrayal of safe sexual practices and prioritising the enjoyment of all involved</p>
<p>• keep in mind the difference between fantasies and real sexual encounters when watching porn. Remember that what you see online may not translate to real life</p>
<p>• keep a close eye on your porn consumption. If you sense it’s becoming overwhelming, or impacting your daily life or sexual experiences, don’t be ashamed. Seek support from a professional, such as a sex therapist.</p>
<hr>
<figure class="align-left ">
<img alt="" src="https://images.theconversation.com/files/536131/original/file-20230706-17-460x2d.png?ixlib=rb-1.1.0&q=45&auto=format&w=237&fit=clip" srcset="https://images.theconversation.com/files/536131/original/file-20230706-17-460x2d.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=600&fit=crop&dpr=1 600w, https://images.theconversation.com/files/536131/original/file-20230706-17-460x2d.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=600&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/536131/original/file-20230706-17-460x2d.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=600&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/536131/original/file-20230706-17-460x2d.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=754&fit=crop&dpr=1 754w, https://images.theconversation.com/files/536131/original/file-20230706-17-460x2d.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=754&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/536131/original/file-20230706-17-460x2d.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=754&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption"></span>
</figcaption>
</figure>
<p><em>Looking for something good? Cut through the noise with a carefully curated selection of the latest releases, live events and exhibitions, straight to your inbox every fortnight, on Fridays. <a href="https://theconversation.com/uk/newsletters/something-good-156">Sign up here</a>.</em></p>
<hr><img src="https://counter.theconversation.com/content/225036/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Chantal Gautier does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.</span></em></p>People in their 20s and 30s are more likely than any other age group to experiment with AI pornography.Chantal Gautier, Lecturer, Sex and Relationship Therapist, University of WestminsterLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2255292024-03-14T13:28:36Z2024-03-14T13:28:36ZHow news organisations decide whether a photo is ‘too edited’<figure><img src="https://images.theconversation.com/files/581726/original/file-20240313-24-56o3zs.jpg?ixlib=rb-1.1.0&rect=43%2C153%2C4844%2C2601&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/soft-focus-women-photographer-hold-camera-1043596813">MheePanda/Shutterstock</a></span></figcaption></figure><p>In the era of artificial intelligence and accessible photo editing, you can’t believe everything you see online. One exception, of course, is (usually) if it’s published by a reputable news source. </p>
<p>The foundation of photojournalism lies in its ability to present reality in an
authentic and unaltered manner. Digital manipulation poses a significant threat to this core principle, undermining the credibility and trustworthiness of the images distributed by photo agencies. The controversy around a retouched family photograph of the Princess of Wales and her children was a rare glimpse into how publishers deal with this issue.</p>
<p>Agencies such as Getty Images and PA Images play a crucial role in delivering accurate and reliable photographs to the public. These organisations adhere to strict <a href="https://www.ap.org/about/news-values-and-principles/telling-the-story/visuals">codes of conduct</a> designed to ensure the integrity of the images they distribute. If an image is accepted but later found to violate these guidelines, it is given a “kill order”. It sounds dramatic, but this instantly stops the distribution.</p>
<p>The main reason why photo agencies cannot accept digitally manipulated imagery is the potential distortion of truth. Manipulated photos can present a skewed version of reality, misinforming the public and compromising the public’s trust in them. Many a <a href="https://www.poynter.org/reporting-editing/2003/l-a-times-photographer-fired-over-altered-image/">photographer</a> has been <a href="https://www.ap.org/ap-in-the-news/2014/ap-severs-ties-with-photographer-who-altered-work">fired</a> for <a href="https://archive.nytimes.com/thelede.blogs.nytimes.com/2008/07/10/in-an-iranian-image-a-missile-too-many/">violating</a> this <a href="https://www.theguardian.com/media/2015/mar/05/world-press-photo-award-withdrawn-over-violation-of-rules">trust</a>. </p>
<p>Photojournalism is a powerful tool for documenting and bearing witness to events around the world. Authenticity is paramount. Even family portraits of public figures become historical documents. </p>
<p>There is a grey area around portraits in the ethical discussion. They can be staged or directed – the photographer will guide and position people. But there is still a requirement in the press to avoid any retouching. That said, in areas such as fashion and celebrity outlets where airbrushing is common, those guidelines are looser.</p>
<p>Photo agencies have their own standards about what level of editing is acceptable. <a href="https://www.afp.com/communication/chartes/12_april_2016_afp_ethic_final.pdf">AFP</a> says that photos and videos “must not be staged, manipulated or edited to give a misleading or false picture of events”. <a href="https://contributors.gettyimages.com/img/articles/downloads/NHLI%20MANIPULATION.pdf">Getty allows</a> for some minor changes such as colour adjustment or removal of red eye or of dust from a dirty lens, but prohibits “extreme” colour or light adjustments.</p>
<p>Several agencies decided to <a href="https://apnews.com/article/kate-princess-photo-surgery-ca91acf667c87c6c70a7838347d6d4fb">retract the photo</a> of the royals because it did not meet their standards. This does not mean the photo was AI-generated or fake, only that it does not meet the strict level of acceptable editing.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/princess-of-wales-photo-controversy-shows-weve-been-thinking-about-edited-images-the-wrong-way-225521">Princess of Wales photo controversy shows we've been thinking about edited images the wrong way</a>
</strong>
</em>
</p>
<hr>
<h2>Changing tech, changing guidelines</h2>
<p>As new technology such as generative AI (which can create photos or videos from a prompt) makes photo editing and creating fake images easier, press agencies are starting to discuss <a href="https://blog.ap.org/standards-around-generative-ai">how to handle it</a>. The Associated Press states:</p>
<blockquote>
<p>We will refrain from transmitting any AI-generated images that are suspected or proven to be false depictions of reality. However, if an AI-generated illustration or work of art is the subject of a news story, it may be used as long as it is clearly labelled as such in the caption.</p>
</blockquote>
<p>News organisations are also experimenting with AI-generated text, and <a href="https://www.niemanlab.org/2023/07/writing-guidelines-for-the-role-of-ai-in-your-newsroom-here-are-some-er-guidelines-for-that/">developing guidelines</a> for this. They tend to focus on transparency, making clear to readers when artificially generated content is being used.</p>
<figure class="align-center ">
<img alt="A man sits at a desktop computer and edits a photo of a model on an editing programme like Photoshop" src="https://images.theconversation.com/files/581733/original/file-20240313-24-o4ujmq.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/581733/original/file-20240313-24-o4ujmq.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=338&fit=crop&dpr=1 600w, https://images.theconversation.com/files/581733/original/file-20240313-24-o4ujmq.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=338&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/581733/original/file-20240313-24-o4ujmq.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=338&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/581733/original/file-20240313-24-o4ujmq.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=424&fit=crop&dpr=1 754w, https://images.theconversation.com/files/581733/original/file-20240313-24-o4ujmq.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=424&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/581733/original/file-20240313-24-o4ujmq.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=424&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">Some retouching is accepted in photojournalism and fashion photography.</span>
<span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/professional-photographer-sitting-his-desk-uses-1599273019">Gorodenkoff/Shutterstock</a></span>
</figcaption>
</figure>
<p>World Press Photo (WPP), an organisation known for its annual photojournalism contest, provides explicit guidelines for submission, <a href="https://www.worldpressphoto.org/contest/2024/verification-process#:%7E:text=Entry%20rule%2017%20states%20photographs,to%20the%20content%20of%20a">updated annually</a>. Photo agencies often align themselves with these principles, recognising the importance of a universal standard for truthfulness in visual reporting.</p>
<p>Due to pressure from photographers and artists who work in more conceptual photography, WPP has added an “open format” category. This welcomes “innovative techniques, non-traditional modes of presentation, and new approaches to storytelling”. The contest organisers considered allowing AI-generated images in 2023, but <a href="https://www.worldpressphoto.org/news/2023/could-an-ai-image-win-our-contest">backtracked</a> after outrage from many photojournalists. </p>
<p>The rise of advanced editing tools and software has made it harder to distinguish between authentic and manipulated images. Fully embracing manipulated imagery in a photojournalism contest would be a risk to the industry’s credibility, at a time when trust in journalism is already at risk.</p><img src="https://counter.theconversation.com/content/225529/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Andrew Pearsall does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.</span></em></p>Publishers are starting to develop guidelines around the use of generative AI in photos.Andrew Pearsall, Senior Lecturer in Photojournalism, University of South WalesLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2249842024-03-14T13:28:28Z2024-03-14T13:28:28ZGhostbots: AI versions of deceased loved ones could be a serious threat to mental health<figure><img src="https://images.theconversation.com/files/580740/original/file-20240308-29-sis8wp.jpg?ixlib=rb-1.1.0&rect=5%2C26%2C3565%2C2350&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/female-face-matrix-digital-numbers-artifical-2268966863">Alena Ivochkina/Shutterstock</a></span></figcaption></figure><p>We all experience loss and grief. Imagine, though, that you don’t need to say goodbye to your loved ones. That you can recreate them virtually so you can have conversations and find out how they’re feeling. </p>
<p>For Kim Kardashian’s fortieth birthday, her then husband, Kanye West, gave her a hologram of her <a href="https://www.bbc.co.uk/news/entertainment-arts-54731382">dead father</a>, Robert Kardashian. Reportedly, Kim Kardashian reacted with disbelief and joy to the <a href="https://www.theguardian.com/lifeandstyle/2020/oct/30/robert-kardashian-resurrected-as-a-hologram-for-kim-kardashian-wests-birthday">virtual appearance</a> of her father at her birthday party. Being able to see a long-dead, much missed loved one, moving and talking again might offer comfort to those left behind. </p>
<p>After all, resurrecting a deceased loved one might seem miraculous – and possibly more than a little creepy – but what’s the impact on our health? Are AI ghosts a help or hindrance to the grieving process? </p>
<p>As a psychotherapist researching how AI technology can be used to enhance therapeutic interventions, I’m intrigued by the advent of ghostbots. But I’m also more than a little concerned about the potential effects of this technology on the mental health of those using it, especially those who are grieving. Resurrecting dead people as avatars has the potential to cause more harm than good, perpetuating even more confusion, stress, depression, paranoia and, in some cases, psychosis.</p>
<p>Recent developments in artificial intelligence (AI) have led to the creation of ChatGPT and other chatbots that can allow users to have sophisticated human like conversations.</p>
<p>Using deep fake technology, AI software can create an <a href="https://www.sciencedirect.com/science/article/pii/S0267364924000104">interactive virtual representation</a> of a deceased person by using their <a href="https://wired.me/technology/artificial-intelligence/why-scientists-are-building-ai-powered-digital-imprints-of-the-dead/">digital content</a> such as photographs, emails, and videos. </p>
<p>Some of these creations were just themes in science fiction fantasy only a few years ago but now they are a scientific reality. </p>
<h2>Help or hindrance?</h2>
<p><a href="https://link.springer.com/article/10.1007/s12124-022-09679-3">Digital ghosts</a> could <a href="https://www.technologyreview.com/2022/10/18/1061320/digital-clones-of-dead-people/">be a comfort</a> to the bereaved by helping them to reconnect with lost loved ones. They could provide an opportunity for the user to say some things or ask questions they never got a chance to when the now deceased person was alive. </p>
<p>But the ghostbots’ uncanny resemblance to a lost loved one <a href="https://www.newscientist.com/article/2416079-resurrecting-loved-ones-as-ai-ghosts-could-harm-your-mental-health/">may not be</a> as positive as it sounds. <a href="https://link.springer.com/article/10.1007/s10676-024-09744-y">Research suggests</a> that deathbots should be used only as a temporary <a href="https://link.springer.com/article/10.1007/s12124-022-09679-3">aid to mourning</a> to avoid potentially harmful emotional dependence on the technology.</p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/IJeqTUG75gA?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
</figure>
<p>AI ghosts could be harmful for people’s mental health by interfering with the <a href="https://www.newscientist.com/article/mg26034650-700-how-ai-avatars-of-the-deceased-could-transform-the-way-we-grieve/">grief process</a>. </p>
<p>Grief takes time and there are many <a href="https://www.medicalnewstoday.com/articles/grieving-process#:%7E:text=They%20include%20shock%2C%20denial%2C%20anger,them%20cope%20in%20various%20ways.">different stages</a> that can take place over many years. When newly bereaved, those experiencing grief might think of their deceased loved one frequently. They might freshly recall old memories and it is quite common for a grieving person <a href="https://pubmed.ncbi.nlm.nih.gov/23449603/">to dream</a> more intensely about their lost loved one. </p>
<p>The psychoanalyst <a href="https://tidsskriftet.no/en/2020/03/essay/dynamics-grief-and-melancholia">Sigmund Freud</a> was concerned with how human beings respond to the experience of loss. He pointed out potential added difficulties for those grieving if there’s negativity surrounding a death. </p>
<p>For example, if a person had ambivalent feelings towards someone and they died, the person could be left with a sense of guilt. Or if a person died in <a href="https://www.frontiersin.org/journals/psychiatry/articles/10.3389/fpsyt.2020.00749/full">horrific circumstances</a> such as a murder, a grieving person might find it more difficult to accept it this. </p>
<p>Freud referred to this as “melancholia”, but it can also be referred to as <a href="https://www.tandfonline.com/doi/abs/10.1080/15524256.2020.1745726">“complicated grief”</a>. In some extreme cases, a person may experience apparitions <a href="https://journals.sagepub.com/doi/full/10.1177/1363461520962887">and hallucinate</a> that they see the dead person and begin to believe they are alive. AI ghostbots could further traumatise someone experiencing complicated grief and may exacerbate associated problems such as hallucinations.</p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/Q8DIBNkghs8?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
</figure>
<h2>Chatbot horror</h2>
<p>There are also risks that these ghost-bots could say harmful things or give bad advice to someone in mourning. Similar generative software such as ChatGPT chatbots are already widely criticised for <a href="https://www.nytimes.com/2023/02/08/technology/ai-chatbots-disinformation.html">giving misinformation</a> to users. </p>
<p>Imagine if the AI technology went rogue and started to make <a href="https://www.nytimes.com/2023/02/16/technology/bing-chatbot-microsoft-chatgpt.html">inappropriate remarks</a> to the user – a situation experienced by journalist <a href="https://www.nytimes.com/2023/02/16/technology/bing-chatbot-transcript.html">Kevin Roose</a> in 2023 when a <a href="https://www.nytimes.com/2023/12/17/insider/ai-chatbots-humans-hallucinate.html#:%7E:text=On%20Valentine's%20Day%20this%20year,him%20to%20leave%20his%20wife.">Bing chatbot</a> tried to get him to leave his wife. It would be very hurtful if a deceased father was conjured up as an AI ghost by a son or daughter to hear comments that they weren’t loved or liked or weren’t their father’s favourite. </p>
<p>Or, in a more extreme scenario, if the ghostbot suggested the user join them in death or they should kill or harm someone. This may sound like a plot from a horror film but it’s not so far fetched. In 2023, the UK’s <a href="https://www.bbc.co.uk/news/uk-politics-66224052">Labour party</a> outlined a law to prevent the training of AI to incite violence. </p>
<p>This was a response to the attempted assassination of the Queen earlier in the year by a man who was encouraged by his chatbot girlfriend, with whom he had an “<a href="https://www.bbc.co.uk/news/uk-england-berkshire-66123122">emotional and sexual</a>” relationship.</p>
<p>The creators of ChatGPT currently acknowledge that the software makes errors and is still <a href="https://www.scientificamerican.com/article/ai-platforms-like-chatgpt-are-easy-to-use-but-also-potentially-dangerous/">not fully reliable</a> because it fabricates information. Who knows how a person’s texts, emails or videos will be interpreted and what content will be generated by this AI technology? </p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/s2FJRbRsBBY?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
</figure>
<p>In any event, it appears that no matter how far this technology advances, there will be a need for considerable oversight and human supervision.</p>
<h2>Forgetting is healthy</h2>
<p>This latest tech says a lot about our digital culture of infinite possibilities with no limits.</p>
<p>Data can be stored on the cloud indefinitely and everything is retrievable and nothing truly deleted or destroyed. Forgetting is an important element of healthy grief but in order to forget, people will need to find new and meaningful ways of remembering the deceased person.</p>
<p>Anniversaries play a key role in helping those who are mourning to not only remember lost loved ones, but they are also opportunities to <a href="https://scholarlypublishingcollective.org/psup/speculative-philosophy/article-abstract/34/3/284/196788/Grief-Phantoms-and-Re-membering-Loss">represent the loss</a> in new ways. Rituals and symbols can mark the end of something that can allow humans to properly remember in order to properly forget.</p><img src="https://counter.theconversation.com/content/224984/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Nigel Mulligan does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.</span></em></p>AI ghosts, the recreation of digital versions of the dead, may sound like a wonderful idea to those dealing with the pain of loss but this technology could seriously disrupt the grieving processNigel Mulligan, Assistant Professor in Psychotherapy, School of Nursing, Psychotherapy and Community Health, Dublin City UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2240532024-03-13T12:38:21Z2024-03-13T12:38:21ZHow AI is shaping the music listening habits of Gen Z<figure><img src="https://images.theconversation.com/files/581018/original/file-20240311-22-us2x6s.jpg?ixlib=rb-1.1.0&rect=40%2C16%2C5339%2C3565&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Listening to music from a device creates a protective bubble that can counteract a lack of personal space at school or home.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/news-photo/physeline-michel-a-member-of-the-haitian-female-soccer-team-news-photo/1213010698?adppopup=true">Pierre Michel Jean/AFP via Getty Images</a></span></figcaption></figure><p>For four years, we’ve been teaching a class on music and the mind. We’ve asked the students at the start of each semester to complete a short, informal survey on their music education and favorite songs and artists. </p>
<p>Our students’ musical education backgrounds always range from none to more than a decade of lessons and ensembles. But we’ve watched the list of favorite songs and artists get longer and more varied each year. When we ask the entire group about certain songs, it is often the case that no one, save for the person who included it, has heard it.</p>
<p>The findings from these informal classroom surveys are consistent with recent research showing diverse and eclectic musical preferences among adolescents. In a study on <a href="https://doi.org/10.3389/fpsyg.2022.1056542">the listening habits of Los Angeles middle school students</a>, we found that they appreciate artists representing a range of genres, from the <a href="https://theconversation.com/how-the-korean-boy-band-bts-toppled-asian-stereotypes-and-took-america-by-storm-97596">K-pop supergroup BTS</a> to the heavy metal band <a href="https://systemofadown.com/">System of a Down</a> <a href="https://theconversation.com/how-a-team-of-musicologists-and-computer-scientists-completed-beethovens-unfinished-10th-symphony-168160">to Beethoven</a>.</p>
<p>But what happens when, as we’ve observed, young people don’t know what their peers are listening to? And does it matter that teens aren’t necessarily choosing the music they’re using to understand themselves and the world, let alone that no humans are selecting songs they’re exposed to?</p>
<h2>A shared soundscape goes private</h2>
<p>For centuries, the only way to experience music was to see it live – at small, private performances, in community gatherings or in large concert halls.</p>
<p>Radios and record players transformed how people interacted with music. But because these devices were initially stationary, there was still a social element to listening. You might gather in a friend’s basement to hear hits on the radio, throw a listening party when a new album was released, make a mixtape for your beau or belt out a favorite song on the car radio with your best friend. </p>
<p>Introduced in 1979, <a href="https://www.smithsonianmag.com/innovation/walkman-invention-40-years-ago-launched-cultural-revolution-180972552/">the Sony Walkman</a> marked another major turning point in how people listen to music. It became a lot easier for music to be a deeply private and personal experience – even more so with the introduction of the iPod and, later, smartphones. </p>
<figure class="align-center ">
<img alt="Photo from the 1950s of three teenage girls relaxing on a carpet listening to records." src="https://images.theconversation.com/files/581077/original/file-20240311-89474-dtltg3.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/581077/original/file-20240311-89474-dtltg3.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=475&fit=crop&dpr=1 600w, https://images.theconversation.com/files/581077/original/file-20240311-89474-dtltg3.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=475&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/581077/original/file-20240311-89474-dtltg3.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=475&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/581077/original/file-20240311-89474-dtltg3.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=596&fit=crop&dpr=1 754w, https://images.theconversation.com/files/581077/original/file-20240311-89474-dtltg3.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=596&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/581077/original/file-20240311-89474-dtltg3.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=596&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">Friends used to get together to listen to music far more often than they do today.</span>
<span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/news-photo/1950s-three-teen-girls-talking-listening-to-music-playing-news-photo/563940019?adppopup=true">H. Armstrong Roberts/ClassicStock via Getty Images</a></span>
</figcaption>
</figure>
<p>Listening to music this way isn’t always about what’s pulsing through your headphones. It can also <a href="https://doi.org/10.3389/fpsyg.2019.02911">cultivate agency</a>: No matter where you are, you are your own DJ, controlling what gets played and when. And if you choose to keep it private, no one can hear it but you. </p>
<p>Particularly for adolescents, this is a big deal. <a href="https://oshkoshnorthstar.org/4858/columns/why-are-teenagers-dependent-on-headphones/">It creates a protective bubble</a> that may counteract a lack of personal space at school or at home.</p>
<p>Young people listen to a lot of music <a href="https://doi.org/10.4103%2Fnah.NAH_65_16">throughout the day</a>, whether it’s while doing homework, training for sports, eating or even sleeping. There’s an <a href="http://st.markgroves.de/wp-content/uploads/2016/10/The-role-of-music-2.pdf">element of mood regulation at play</a>: Songs can divert unpleasant emotions or elicit positive ones, and also encourage reflection during difficult experiences.</p>
<h2>I got ‘algo-rhythm’</h2>
<p>Making a playlist used to mean playing tapes and recording individual songs onto another tape, or waiting for the radio to play a song, hitting “record” on your cassette player to capture it, song by song, <a href="https://www.wired.com/story/history-of-mixtapes-future/">until you had a mixtape of your favorite tunes</a>.</p>
<p>Now, listening <a href="https://www.ifpi.org/wp-content/uploads/2023/12/IFPI-Engaging-With-Music-2023_full-report.pdf">often happens via streaming</a>, where artificial intelligence and social media platforms team up to suggest playlists for you. </p>
<p>While you explore and share music on social media, <a href="https://neemz.medium.com/the-inner-workings-of-spotifys-ai-powered-music-recommendations-how-spotify-shapes-your-playlist-a10a9148ee8d">AI tracks the activity</a> and compares it to data from other listeners; in this way, it hones its predictions about what you might like to hear in the future. </p>
<p>AI is being put to work to know not only what a user wants to hear, but also to predict the next big hit that everyone will listen to. Until recently, AI’s power for predicting hits relied largely on song characteristics like <a href="https://newsroom.spotify.com/2022-11-30/learn-about-those-music-genres-you-may-not-have-heard-of/">bounciness, positiveness and danceability</a>, and hovered at around 50% accuracy.</p>
<p><a href="https://doi.org/10.3389/frai.2023.1154663">Other studies</a> have analyzed physiological responses to music, like heart rate, which can be gleaned from the biodata on teen’s smartwatches, <a href="https://www.scientificamerican.com/podcast/episode/heres-how-ai-can-predict-hit-songs-with-frightening-accuracy/">to predict top hits</a>.</p>
<p>These studies add to existing concerns about the mining of personal information and data, and there have long been fears that <a href="https://encyclopedia.pub/entry/35791">AI can’t be trusted and will end up manipulating people</a>. When it comes to the way AI influences your listening habits, you might wonder whether you like a song because you truly like it, or whether you only enjoy it because AI has fed you enough similar songs that familiarity has bred appreciation.</p>
<p>Some listeners feel that algorithmic curation causes them to be <a href="https://theconversation.com/i-almost-feel-like-stuck-in-a-rut-how-streaming-services-changed-the-way-we-listen-to-music-219967">stuck in a listening rut</a>. Their playlists are populated with songs and artists they’ve never heard of before, yet they all sound eerily similar.</p>
<h2>The upside to AI</h2>
<p>In the past, being in a listening rut was something a teenager may not have even noticed.</p>
<p>Exposed to a steady diet of the same songs regularly playing on the radio – and later, <a href="https://www.npr.org/templates/story/story.php?storyId=96869060">on MTV</a> and VH1 – adolescents’ musical consumption was dominated by the “Top-40” artists. Their palettes were sculpted by a widely shared, if perhaps narrow, repertoire of musical knowledge.</p>
<figure class="align-center ">
<img alt="Two young women and one young man pose in front of screaming fans." src="https://images.theconversation.com/files/581003/original/file-20240311-18-gp310p.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/581003/original/file-20240311-18-gp310p.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=391&fit=crop&dpr=1 600w, https://images.theconversation.com/files/581003/original/file-20240311-18-gp310p.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=391&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/581003/original/file-20240311-18-gp310p.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=391&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/581003/original/file-20240311-18-gp310p.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=491&fit=crop&dpr=1 754w, https://images.theconversation.com/files/581003/original/file-20240311-18-gp310p.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=491&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/581003/original/file-20240311-18-gp310p.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=491&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">Jennifer Lopez, Justin Timberlake and Halle Berry appear at MTV Studios in New York’s Times Square for a taping of ‘TRL’ during the network’s ‘Spankin’ New Music Week’ in 2002.</span>
<span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/news-photo/jennifer-lopez-justin-timberlake-and-halle-berry-during-news-photo/107285046?adppopup=true">KMazur/WireImage via Getty Images</a></span>
</figcaption>
</figure>
<p>AI-generated playlists have disrupted this, and the two of us don’t see that as necessarily a bad thing. A stunning range of music is available to young people, and no longer do radio DJs, ratings and record companies serve as gatekeepers. </p>
<p>Spotify currently <a href="https://gist.github.com/andytlr/4104c667a62d8145aa3a">lists thousands of genres</a> and creates more each year so that, <a href="https://newsroom.spotify.com/2022-11-30/learn-about-those-music-genres-you-may-not-have-heard-of/">as the company explains</a>, they are more “recognizable, representative, and holistic to our listeners and communities.”</p>
<p>Like receiving a cherished gift you never knew you wanted, young people can be exposed to great music – with its accompanying cultural traditions – that they would be less likely to have discovered on their own, whether it’s <a href="https://medium.com/@khushibagwar092/indian-90s-pop-culture-861ad6250d3d">Indian pop music</a>, <a href="https://www.masterclass.com/articles/japanese-rock-music-guide">Japanese rock</a> or <a href="https://www.africanmusiclibrary.org/genre/Juju">Afro-juju</a>, a style of Nigerian popular music.</p>
<p>If teens think their AI-influenced playlists are dull, they still have the ability to search for new music. Just because algorithms and AI can suggest songs, it doesn’t preclude listeners from researching and discovering music on their own, or sharing playlists with friends and relatives.</p>
<p>Anything that exists, they can find. The store is always open.</p>
<h2>Identity, community and music</h2>
<p>Back to our college class: We noticed little overlap among the students. But instead of consuming only from a menu of industry megastars, our students showed a willingness to listen to a variety of genres and subgenres that AI will offer up. </p>
<p>When asked to reveal the most recent song or piece that they had listened to on a specific week, 6% had listened to R&B singer SZA, 2% to singer Renée Rapp, 2% to pop sensation Taylor Swift and 2% to pop rockers The 1975.</p>
<p>The remaining 80-plus selections featured a panoply of genres: <a href="https://electronicmusic.fandom.com/wiki/Computer_music">computer music</a>, rock, pop, rap, country, reggaeton, film music, heavy metal, indie and Latin ballads. </p>
<p>As young people transition from childhood to adulthood, <a href="https://www.youtube.com/watch?v=REP6ZNV0OR4">two seemingly opposing processes become paramount</a>: forming a unique identity, while at the same time becoming part of a community. Music listening and preferences play an important role in <a href="https://www.youtube.com/watch?v=mblJUqyizwg">this process</a>.</p>
<p>AI-generated playlists have the potential to challenge this transition. </p>
<p>So does AI make it easier to differentiate the self, but harder to bond with others? Or does it, instead, offer a broader spectrum for self-exploration and communal connection? </p>
<p>The truth is, no one really knows. </p>
<p><a href="https://theconversation.com/chatgpt-is-confronting-but-humans-have-always-adapted-to-new-technology-ask-the-mesopotamians-who-invented-writing-199184">Fears of new technologies</a> are commonplace. For example, as scheduled network TV <a href="https://www.wsj.com/articles/the-decade-in-television-betteryet-more-atomizedthan-ever-11576620996">fell out of favor</a>, a lot of common ground for discussion and connection disappeared with it. Will 50 million Americans ever again tune in to watch the series finale of a sitcom, <a href="https://theconversation.com/4-reasons-why-well-never-see-another-show-like-friends-123411">as they did for “Friends” in 2004</a>?</p>
<p>If AI is, indeed, contributing to the transformation of adolescents’ communal listening experiences, then AI playlists are more than just a convenient way to discover your next workout tune. They are a revolution worth paying attention to.</p><img src="https://counter.theconversation.com/content/224053/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Beatriz Ilari received funding from The Fender Play Foundation to carry out the study with Angeleno adolescents that is mentioned in the article. </span></em></p><p class="fine-print"><em><span>Lynne Snyder does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.</span></em></p>In the past, adolescents’ musical palettes were dominated by the Top-40 artists, creating a widely shared – if perhaps narrow – repertoire of musical knowledge.Beatriz Ilari, Professor of Music Teaching and Learning, University of Southern CaliforniaLynne Snyder, Doctor of Musical Arts Student, University of Southern CaliforniaLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2240442024-03-13T12:28:20Z2024-03-13T12:28:20ZRobo-advisers are here – the pros and cons of using AI in investing<figure><img src="https://images.theconversation.com/files/580679/original/file-20240308-28-55toe3.jpg?ixlib=rb-1.1.0&rect=59%2C0%2C7951%2C4345&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">shutterstock</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/smart-businessman-hand-close-nft-financial-2074315681">thinkhubstudio/Shutterstock</a></span></figcaption></figure><p>Artificial intelligence (AI) is <a href="https://www.ft.com/content/6766a3bd-1cec-4e88-9f51-5ed93b39528c">shaking up</a> the way we invest our money. Gone are the days when complex tools were reserved for the wealthy or financial institutions. </p>
<p>AI-powered <a href="https://www.investopedia.com/best-robo-advisors-4693125">robo-advisers</a>, such as <a href="https://www.betterment.com/">Betterment</a> and <a href="https://investor.vanguard.com/advice/robo-advisor">Vanguard</a> in the US, and finance app <a href="https://www.revolut.com/en-HU/news/revolut_launches_robo_advisor_in_eea_to_automate_investing/">Revolut</a> in Europe, are now democratising investment. These tools are making professional financial insight and portfolio management available to everyone. But although there are plenty of advantages to using robo-advisers, there are downsides too. </p>
<p>Since the 1990s, <a href="https://arxiv.org/pdf/2104.05413.pdf">AI’s role</a> in this sector was typically confined to algorithmic trading and quantitative strategies. These rely on advanced mathematical models to predict stock market movements and trade at lightning speed, far exceeding the capabilities of human traders. </p>
<p>But that laid the groundwork for more advanced applications. And AI has now <a href="https://www.weforum.org/agenda/2017/09/robots-could-plan-your-retirement-financial-advice/">evolved</a> to handle data analysis, predict trends and personalise investment strategies. Unlike traditional investment tools, robo-advisers are more <a href="https://www2.deloitte.com/us/en/insights/industry/financial-services/financial-services-industry-predictions/2023/democratize-financial-services.html">accessible</a>, making them ideal for a new generation of investors. </p>
<p>A survey published in 2023 showed that there has been a particular <a href="https://www.investopedia.com/study-affluent-millennials-are-warming-up-to-robo-advisors-4770577">surge</a> in young people using robo-advisers. Some 31% of gen Zs (born after 2000) and 20% of millennials (born between 1980 and 2000) are using robo-advisers. </p>
<p>Another <a href="https://www.magnifymoney.com/news/robo-advisor-survey/">survey</a> from 2022 found that 63% of US consumers were open to using a robo-adviser to manage their investments. In fact, projections indicate that assets managed by robo-advisers will reach <a href="https://www.statista.com/outlook/fmo/wealth-management/digital-investment/robo-advisors/worldwide">US$1.8 trillion</a> (£1.4 trillion) globally in 2024. </p>
<p>This trend reflects not only changing investor preferences but also how the financial industry is adapting to technology.</p>
<h2>Tailored advice</h2>
<p>AI can <a href="https://www.ftadviser.com/your-industry/2023/07/17/can-generative-ai-truly-replace-a-financial-adviser/">tailor</a> investment advice to a person’s preferences. For example, for investors who want to prioritise ethical investing in environmental, social and governance stocks, AI can tailor a strategy without the need to pay for a financial adviser. </p>
<p>AI can <a href="https://www.sciencedirect.com/science/article/pii/S0275531923000077">analyse</a> news and social media to understand market trends and predict potential movements, offering insights into potential market movements. Portfolios built by robo-advisers may also be <a href="https://onlinelibrary.wiley.com/doi/full/10.1111/poms.14029">more resilient during market downturns</a>, effectively managing risk and protecting investments.</p>
<p>Robo-advisers can offer certain <a href="https://www.ft.com/content/6694bb4a-a585-496a-b7f3-d1841984f9b3">features</a> like reduced investment account minimums and lower fees, which make services more accessible than in the past. Other features such as <a href="https://corporatefinanceinstitute.com/resources/wealth-management/robo-advisors/">tax-loss harvesting</a>, a strategy of selling assets at a loss to reduce taxes, and <a href="https://corporatefinanceinstitute.com/resources/wealth-management/robo-advisors/">periodic rebalancing</a>, which involves adjusting the proportions of different types of investments, make professional investment advice accessible to a wider audience.</p>
<p>These types of innovations are particularly beneficial for people in underserved communities or with limited financial resources. This has the <a href="https://www.brookings.edu/articles/robo-advice-an-effective-tool-to-reduce-inequalities/">potential</a> to improve financial literacy through empowering people to make better financial decisions. </p>
<h2>AI’s multifaced role</h2>
<p>AI’s impact on investment fund management goes way beyond robo-advisers, however. Fund managers are using AI algorithms in a variety of ways. </p>
<p>In terms of data analysis, AI can sift through vast amounts of market data and historical trends to identify <a href="https://doi.org/10.1016/j.frl.2022.102941">ideal assets</a> and adjust portfolios in real time as markets fluctuate. AI is also used to <a href="https://www.sciencedirect.com/science/article/pii/S0378426621002466">improve risk management</a> by analysing complex data and making sophisticated decisions. </p>
<p>By using AI in this way, <a href="https://doi.org/10.1016/j.jedc.2022.104438">traders</a> can react and make faster decisions, which maximises efficiency. Other mundane tasks like <a href="https://ieeexplore.ieee.org/document/9315986">compliance monitoring</a> are increasingly automated by AI. This frees fund managers up to focus on more strategic decisions. </p>
<figure class="align-center ">
<img alt="A close up of a pair of hands holding a mobile phone with pound coins superimposed onto the foreground." src="https://images.theconversation.com/files/580727/original/file-20240308-24-xg6lqw.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/580727/original/file-20240308-24-xg6lqw.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=350&fit=crop&dpr=1 600w, https://images.theconversation.com/files/580727/original/file-20240308-24-xg6lqw.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=350&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/580727/original/file-20240308-24-xg6lqw.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=350&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/580727/original/file-20240308-24-xg6lqw.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=440&fit=crop&dpr=1 754w, https://images.theconversation.com/files/580727/original/file-20240308-24-xg6lqw.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=440&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/580727/original/file-20240308-24-xg6lqw.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=440&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">While AI is democratising investing, that comes with challenges.</span>
<span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/double-exposure-uk-stock-graphic-close-792232471">Loch Earn/Shutterstock</a></span>
</figcaption>
</figure>
<h2>What are the disadvantages?</h2>
<p>One of the biggest concerns regarding AI in this sector is based on how having easy access to advanced investment tools may lead some people to overestimate their abilities and take too many financial risks. The sophisticated algorithms used by robo-investors can be opaque, which makes it <a href="https://www.lseg.com/en/insights/data-analytics/how-might-ai-impact-investment-management">difficult</a> for some investors to fully understand the potential risks involved. </p>
<p>Another concern is how the evolution of robo-advisers has outpaced the implementation of <a href="https://fastercapital.com/content/Regulatory-Compliance-in-B2B-Robo-Advisors--Navigating-the-Legal-Landscape.html#Challenges-and-Opportunities">laws and regulations</a>. That could expose investors to financial risks and a lack of legal protection. This is an issue yet to be adequately addressed by financial authorities. </p>
<p>Looking ahead, the future of investment probably lies in a hybrid model. Combining the precision and efficiency of AI with the experience and oversight of human investors is vital.</p>
<p>Ensuring that information is accessible and transparent will be crucial for <a href="https://www.turing.ac.uk/sites/default/files/2021-06/ati_ai_in_financial_services_lores.pdf">fostering</a> a more informed and responsible investment landscape. By harnessing the power of AI responsibly, we can create a financial future that benefits everyone.</p><img src="https://counter.theconversation.com/content/224044/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>The authors do not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.</span></em></p>Robo-advisers and AI are making investing accessible to everyone, but there are also risks to consider.Laurence Jones, Lecturer in Finance, Bangor UniversityHeather He, Lecturer in Data Science/Analytics, Bangor UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2250802024-03-12T19:15:10Z2024-03-12T19:15:10ZAI is creating fake legal cases and making its way into real courtrooms, with disastrous results<figure><img src="https://images.theconversation.com/files/581132/original/file-20240312-16-84kk3r.jpg?ixlib=rb-1.1.0&rect=30%2C24%2C4082%2C2713&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/lady-justice-on-digital-background-concept-1044578125">Shutterstock</a></span></figcaption></figure><p>We’ve seen deepfake, explicit images of <a href="https://www.nytimes.com/2024/01/26/arts/music/taylor-swift-ai-fake-images.html">celebrities</a>, created by artificial intelligence (AI). AI has also played a hand in <a href="https://theconversation.com/the-ghosts-of-the-past-pop-music-is-haunted-by-our-anxieties-about-the-future-218555">creating music</a>, <a href="https://theconversation.com/bringing-ai-up-to-speed-autonomous-auto-racing-promises-safer-driverless-cars-on-the-road-214208">driverless race cars</a> and spreading <a href="https://theconversation.com/misinformation-how-fact-checking-journalism-is-evolving-and-having-a-real-impact-on-the-world-218379">misinformation</a>, among other things.</p>
<p>It’s hardly surprising, then, that AI also has a strong impact on our legal systems.</p>
<p>It’s well known that courts must decide disputes based on the law, which is presented by lawyers to the court as part of a client’s case. It’s therefore highly concerning that fake law, invented by AI, is being used in legal disputes. </p>
<p>Not only does this pose issues of legality and ethics, it also threatens to undermine faith and trust in global legal systems.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/lawyers-are-rapidly-embracing-ai-heres-how-to-avoid-an-ethical-disaster-221135">Lawyers are rapidly embracing AI: here's how to avoid an ethical disaster</a>
</strong>
</em>
</p>
<hr>
<h2>How do fake laws come about?</h2>
<p>There is little doubt that generative AI is a powerful tool with transformative potential for society, including many aspects of the legal system. But its use comes with responsibilities and risks.</p>
<p>Lawyers are trained to carefully apply professional knowledge and experience, and are generally not big risk-takers. However, some unwary lawyers (and <a href="https://reason.com/volokh/2023/11/13/self-represented-litigants-use-ai-to-write-briefs-produce-hallucinated-citations/">self-represented</a> litigants) have been caught out by artificial intelligence.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/581133/original/file-20240312-16-rhmkk1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="ChatGPT on a smartphone screen in front of the same website on a laptop screen" src="https://images.theconversation.com/files/581133/original/file-20240312-16-rhmkk1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/581133/original/file-20240312-16-rhmkk1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/581133/original/file-20240312-16-rhmkk1.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/581133/original/file-20240312-16-rhmkk1.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/581133/original/file-20240312-16-rhmkk1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/581133/original/file-20240312-16-rhmkk1.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/581133/original/file-20240312-16-rhmkk1.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">Generative AI tools, like ChatGPT, can provide incorrect information.</span>
<span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/chatgpt-chat-bot-screen-seen-on-2237655785">Shutterstock</a></span>
</figcaption>
</figure>
<p>AI models are trained on massive data sets. When prompted by a user, they can create new content (both text and audiovisual). </p>
<p>Although content generated this way can look very convincing, it can also be inaccurate. This is the result of the AI model attempting to “fill in the gaps” when its training data is inadequate or flawed, and is commonly referred to as “<a href="https://www.csiro.au/en/news/All/Articles/2023/June/humans-and-ai-hallucinate">hallucination</a>”.</p>
<p>In some contexts, generative AI hallucination is not a problem. Indeed, it can be seen as an example of creativity. </p>
<p>But if AI hallucinated or created inaccurate content that is then used in legal processes, that’s a problem – particularly when combined with time pressures on lawyers and a lack of access to legal services for many. </p>
<p>This potent combination can result in carelessness and shortcuts in legal research and document preparation, potentially creating reputational issues for the legal profession and a lack of public trust in the administration of justice.</p>
<h2>It’s happening already</h2>
<p>The best known generative AI “fake case” is the 2023 US case <a href="https://law.justia.com/cases/federal/district-courts/new-york/nysdce/1:2022cv01461/575368/54/">Mata v Avianca</a>, in which lawyers submitted a brief containing fake extracts and case citations to a New York court. The brief was researched using ChatGPT. </p>
<p>The lawyers, unaware that ChatGPT can hallucinate, failed to check that the cases actually existed. The consequences were disastrous. Once the error was uncovered, the court dismissed their client’s case, sanctioned the lawyers for acting in bad faith, fined them and their firm, and exposed their actions to public scrutiny.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/ai-is-everywhere-including-countless-applications-youve-likely-never-heard-of-222985">AI is everywhere – including countless applications you've likely never heard of</a>
</strong>
</em>
</p>
<hr>
<p>Despite adverse publicity, other fake case examples continue to surface. Michael Cohen, Donald Trump’s former lawyer, gave his own lawyer cases generated by Google Bard, another generative AI chatbot. He believed they were real (they were not) and that his lawyer would fact check them (he did not). His lawyer <a href="https://www.reuters.com/legal/ex-trump-fixer-michael-cohen-says-ai-created-fake-cases-court-filing-2023-12-29">included the cases</a> in a brief filed with the US Federal Court.</p>
<p>Fake cases have also surfaced in recent matters in <a href="https://www.cbc.ca/news/canada/british-columbia/lawyer-chatgpt-fake-precedent-1.7126393">Canada</a> and <a href="https://www.legalfutures.co.uk/latest-news/litigant-unwittingly-put-fake-cases-generated-by-ai-before-tribunal">the United Kingdom</a>.</p>
<p>If this trend goes unchecked, how can we ensure that the careless use of generative AI does not undermine the public’s trust in the legal system? Consistent failures by lawyers to exercise due care when using these tools has the potential to mislead and congest the courts, harm clients’ interests, and generally undermine the rule of law.</p>
<h2>What’s being done about it?</h2>
<p>Around the world, legal regulators and courts have responded in various ways. </p>
<p>Several US state bars and courts have issued guidance, opinions or orders on generative AI use, ranging from responsible adoption to an outright ban. </p>
<p>Law societies in the UK and British Columbia, and the courts of New Zealand, have also developed guidelines. </p>
<p>In Australia, the NSW Bar Association has a <a href="https://inbrief.nswbar.asn.au/posts/9e292ee2fc90581f795ff1df0105692d/attachment/NSW%20Bar%20Association%20GPT%20AI%20Language%20Models%20Guidelines.pdf">generative AI guide</a> for barristers. The <a href="https://lsj.com.au/articles/a-solicitors-guide-to-responsible-use-of-artificial-intelligence/">Law Society of NSW</a> and the <a href="https://www.liv.asn.au/Web/Law_Institute_Journal_and_News/Web/LIJ/Year/2023/09September/How_lawyers_are_using_generative_AI.aspx">Law Institute of Victoria</a> have released articles on responsible use in line with solicitors’ conduct rules.</p>
<p>Many lawyers and judges, like the public, will have some understanding of generative AI and can recognise both its limits and benefits. But there are others who may not be as aware. Guidance undoubtedly helps. </p>
<p>But a mandatory approach is needed. Lawyers who use generative AI tools cannot treat it as a substitute for exercising their own judgement and diligence, and must check the accuracy and reliability of the information they receive.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/do-you-trust-ai-to-write-the-news-it-already-is-and-not-without-issues-216909">Do you trust AI to write the news? It already is – and not without issues</a>
</strong>
</em>
</p>
<hr>
<p>In Australia, courts should adopt practice notes or rules that set out expectations when generative AI is used in litigation. Court rules can also guide self-represented litigants, and would communicate to the public that our courts are aware of the problem and are addressing it.</p>
<p>The legal profession could also adopt formal guidance to promote the responsible use of AI by lawyers. At the very least, technology competence should become a requirement of lawyers’ continuing legal education in Australia. </p>
<p>Setting clear requirements for the responsible and ethical use of generative AI by lawyers in Australia will encourage appropriate adoption and shore up public confidence in our lawyers, our courts, and the overall administration of justice in this country.</p><img src="https://counter.theconversation.com/content/225080/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Vicki McNamara is affiliated with the Law Society of NSW (as a member).</span></em></p><p class="fine-print"><em><span>Michael Legg does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.</span></em></p>Generative AI can be a useful tool, but it can also create inaccurate information. Here’s how to safeguard Australian courts against fake cases, like we’ve already seen overseas.Michael Legg, Professor of Law, UNSW SydneyVicki McNamara, Senior Research Associate, Centre for the Future of the Legal Profession, UNSW SydneyLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2243342024-03-12T18:53:02Z2024-03-12T18:53:02ZAncient scrolls are being ‘read’ by machine learning – with human knowledge to detect language and make sense of them<figure><img src="https://images.theconversation.com/files/580263/original/file-20240306-30-3x4aw.png?ixlib=rb-1.1.0&rect=1040%2C0%2C1253%2C379&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">The Vesuvius Challenge incentivizes technological development by inviting researchers to figure out how to ‘read’ ancient papyri excavated from volcanic ash of Mount Vesuvius in Italy. Columns of Greek text retrieved from a portion of a scroll. </span> <span class="attribution"><span class="source">(Vesuvius Challenge)</span>, <span class="license">Author provided</span></span></figcaption></figure><p>A groundbreaking announcement for the recovery of lost ancient literature was recently made. Using a non-invasive method that harnesses <a href="https://mitsloan.mit.edu/ideas-made-to-matter/machine-learning-explained">machine learning</a>, an international trio of scholars retrieved 15 columns of ancient Greek text from within a carbonized papyrus from <a href="https://www.herculaneum.ox.ac.uk/about-us/story-of-herculaneum">Herculaneum</a>, a seaside Roman town eight kilometres southeast of Naples, Italy.</p>
<p>Their achievement earned them a US$700,000 grand prize from the <a href="https://scrollprize.org/">Vesuvius Challenge</a>. The challenge sought to incentivize technological development by inviting public participation in the research. </p>
<p>It emerged from collaboration between computer scientist Brent Seales — who has <a href="https://doi.org/10.48550/arXiv.2304.02084">a long-standing interest</a> in non-invasive <a href="https://www2.cs.uky.edu/dri/the-scroll-from-en-gedi">technologies for studying</a> manuscripts — and technology investors Nat Friedman and Daniel Gross. </p>
<p>While the developments are exciting, technology is only part of the progress of scholarship. The work of reading and analyzing the new Greek and Latin texts recovered from the papyri will fall to human beings.</p>
<figure class="align-center ">
<img alt="Painting showing a mountain with a volcano erupting." src="https://images.theconversation.com/files/580261/original/file-20240306-28-umf4ff.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/580261/original/file-20240306-28-umf4ff.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=404&fit=crop&dpr=1 600w, https://images.theconversation.com/files/580261/original/file-20240306-28-umf4ff.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=404&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/580261/original/file-20240306-28-umf4ff.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=404&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/580261/original/file-20240306-28-umf4ff.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=507&fit=crop&dpr=1 754w, https://images.theconversation.com/files/580261/original/file-20240306-28-umf4ff.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=507&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/580261/original/file-20240306-28-umf4ff.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=507&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">‘An Eruption of Vesuvius,’ by Johan Christian Dahl (1824).</span>
<span class="attribution"><span class="source">(The Metropolitan Museum of Art)</span>, <a class="license" href="http://creativecommons.org/licenses/by/4.0/">CC BY</a></span>
</figcaption>
</figure>
<h2>Buried in ash</h2>
<p>Like Pompeii, <a href="https://www.youtube.com/watch?v=f5b8igA644o">Herculaneum</a> was buried by the catastrophic eruption of Mount Vesuvius in 79 CE. </p>
<p>Much of the ancient town remains underground. But <a href="https://theconversation.com/ai-will-let-us-read-lost-ancient-works-in-the-library-at-herculaneum-for-the-first-time-223583">in 1752</a>, excavation uncovered hundreds of papyrus scrolls in the library of an elaborate Roman villa. The Herculaneum papyri <a href="https://www.herculaneum.ox.ac.uk/research-and-publications/papyri">are the largest surviving example of an</a> intact ancient library preserved in the archaeological record: the library was found as it actually existed in 79 CE. </p>
<p>The precise number of books is unknown, says Michael McOsker, a research fellow in papyrology at University College London, and different methods of estimating give different results. </p>
<h2>Carbonized papyri</h2>
<p>Starved of oxygen, the intense heat of Vesuvius’ <a href="https://www.nationalgeographic.org/encyclopedia/pyroclastic-flow/">pyroclastic flow</a> carbonized (but did not ignite) the papyri. Resembling lumps of coal to the eye, 18th-century excavators did not immediately recognize them as ancient books.</p>
<figure class="align-center ">
<img alt="Three dark grey rectangular objects seen in a box." src="https://images.theconversation.com/files/578769/original/file-20240228-16-sc89zf.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/578769/original/file-20240228-16-sc89zf.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/578769/original/file-20240228-16-sc89zf.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/578769/original/file-20240228-16-sc89zf.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/578769/original/file-20240228-16-sc89zf.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=502&fit=crop&dpr=1 754w, https://images.theconversation.com/files/578769/original/file-20240228-16-sc89zf.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=502&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/578769/original/file-20240228-16-sc89zf.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=502&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">Three unopened papyri from Herculaneum.</span>
<span class="attribution"><span class="source">(Bodleian Libraries/University of Oxford)</span>, <a class="license" href="http://creativecommons.org/licenses/by-nc/4.0/">CC BY-NC</a></span>
</figcaption>
</figure>
<p>The papyri are so brittle that many were destroyed by early attempts to access their texts. Studying them has therefore always required ingenuity. In 1754, a <a href="https://www.smithsonianmag.com/history/buried-ash-vesuvius-scrolls-are-being-read-new-xray-technique-180969358">conservator and priest at the Vatican library</a> devised a machine for slowly unrolling them. </p>
<figure class="align-center ">
<img alt="A dark grey scroll." src="https://images.theconversation.com/files/578666/original/file-20240228-7839-doqnyj.jpg?ixlib=rb-1.1.0&rect=0%2C0%2C7027%2C4995&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/578666/original/file-20240228-7839-doqnyj.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=425&fit=crop&dpr=1 600w, https://images.theconversation.com/files/578666/original/file-20240228-7839-doqnyj.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=425&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/578666/original/file-20240228-7839-doqnyj.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=425&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/578666/original/file-20240228-7839-doqnyj.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=534&fit=crop&dpr=1 754w, https://images.theconversation.com/files/578666/original/file-20240228-7839-doqnyj.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=534&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/578666/original/file-20240228-7839-doqnyj.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=534&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">A portion of an unrolled Herculaneum papyrus.</span>
<span class="attribution"><a class="source" href="https://digital.bodleian.ox.ac.uk/objects/cac4db6a-8af5-4234-%20acb8-4b1ce819ef14">(Bodleian Libraries/University of Oxford)</a>, <a class="license" href="http://creativecommons.org/licenses/by-nc/4.0/">CC BY-NC</a></span>
</figcaption>
</figure>
<p>More recently, <a href="https://www.imaging.org/common/uploaded%20files/pdfs/Papers/2001/PICS-0-251/4625.pdf">multispectral photography</a> has dramatically improved their legibility. But until now, a non-invasive method that would leave the scrolls intact remained out of reach. Its development marks a significant breakthrough.</p>
<p>McOsker notes there are 659 items in the catalogue listed as “not unrolled,” but some of these are parts of scrolls. </p>
<h2>Sparking innovation</h2>
<p>To kick-start the challenge, Seales <a href="https://scrollprize.org/data">made public</a> an array of high-resolution X-ray computed tomography (CT) scans of two scrolls as well as similar scans of detached fragments with visible ink. The latter are essential as a reference point (or “control”) for innovative approaches. </p>
<p>The competition’s design encouraged transparency and collaboration: data published in the pursuit <a href="https://scrollprize.org/winners">of smaller goals</a> benefited all competitors. Additionally, transparency enabled the independent verification of results. Teams coalesced around shared ideas and approaches to the problem.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/ai-will-let-us-read-lost-ancient-works-in-the-library-at-herculaneum-for-the-first-time-223583">AI will let us read 'lost' ancient works in the library at Herculaneum for the first time</a>
</strong>
</em>
</p>
<hr>
<h2>Text mentions music, taste, sight</h2>
<p>The challenge made news in <a href="https://scrollprize.org/firstletters">October</a>, when the first letters were read: πορφυρας (a noun or adjective involving “purple”). </p>
<p>By the end of 2023, the criteria for awarding the grand prize were met: four passages of 140 characters, with 85 per cent of the letters recovered. <a href="https://scrollprize.org/grandprize">A PhD student studying machine learning, an engineer studying computer science and a robotics student</a> were declared
the victors.</p>
<p>According to McOsker, the text they retrieved mentions music twice, as well as the senses of taste and sight. He thinks it is likely a work about sensation and decision-making, in the tradition of <a href="https://plato.stanford.edu/ENTRIES/epicurus/">the philosopher Epicurus (341–270 BCE)</a>. The challenge’s papyrological team is still analyzing it.</p>
<h2>Hundreds of rolls to be studied</h2>
<p>This year brings with it new goals: after five per cent of one scroll was read in 2023, the challenge set a <a href="https://scrollprize.org/2024_prizes#2024-grand-prize">2024 grand prize goal</a> of reading 90 per cent of four scrolls. With hundreds of rolls yet to be studied, the new method of recovering the contents of the Herculaneum papyri is only getting started.</p>
<p>But several obstacles remain. The production of scans at sufficiently high resolution can’t be done via ordinary equipment, but requires access to a facility with a particle accelerator. Access to the right equipment is limited and costly. To date, four scrolls and numerous detached fragments <a href="https://www.diamond.ac.uk">have been processed at a facility</a> near Oxford, England. </p>
<p>Most of the unopened scrolls are housed in Naples, and getting them safely to a facility will be complicated, as will reserving and paying for the beam time required to scan them.</p>
<p>Another limitation is that the technology for unrolling and flattening out a papyrus by virtual means — a process the challenge calls “segmentation” — is slow and expensive. Via current techniques, which involve a fair bit of manual manipulation, fully segmenting one scroll would cost US$1–5 million. Segmentation needs to become much more efficient to avoid a bottleneck.</p>
<h2>Critical minds needed</h2>
<p>Technology is only part of the equation. Essential to the challenge’s work is an international team of papyrologists. Their role is to analyze the model’s output of legible ancient Greek — and in so doing determine which approaches are most effective.</p>
<p>Papyrology is thrilling work, but also challenging and painstaking. It requires mastery of ancient languages and ideas as well as the puzzle-solver’s ability to fill in the inevitable gaps. Papyrology is a niche specialization: in the larger world of classics, papyrologists are rare birds. The number of Herculaneum specialists is even fewer. </p>
<p>For the challenge truly to succeed, we’re going to need critical minds as well as whizbang technology. There’s potentially a fair bit of new ancient philosophy headed our way, but it needs to be pieced together into a coherent text — letter by letter, word by word, sentence by sentence — before it can be studied more widely. That’s going to require scholars.</p><img src="https://counter.theconversation.com/content/224334/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>C. Michael Sampson receives funding from the Social Sciences and Humanities Research Council of Canada for 'the Books of Karanis,' a project that studies fragmentary Greek literature from the Egyptian village Karanis. </span></em></p>However exciting the technological developments may be, the task of reading and analyzing the Greek and Latin texts recovered from the papyri will fall to human beings.C. Michael Sampson, Associate Professor of Classics, University of ManitobaLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2234552024-03-12T02:44:44Z2024-03-12T02:44:44ZArt of the moment: experiencing Marina Abramović and Laurie Anderson at the Adelaide Festival<figure><img src="https://images.theconversation.com/files/581127/original/file-20240311-20-ls9siy.jpg?ixlib=rb-1.1.0&rect=5%2C51%2C3828%2C5702&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><span class="source">Andrew Beveridge/Adelaide Festival</span></span></figcaption></figure><p>Ruth Mackenzie is the new artistic director of the Adelaide Festival of Arts, and for her first Adelaide gig she has brought in two heavyweights: performance artist Marina Abramović and avant-garde artist and musician Laurie Anderson. Both were major events, very much of the moment. </p>
<p>The Marina Abramović Institute’s Takeover featured nine performance artists over four days.</p>
<p>To begin at the beginning, audience members are instructed to arrive at 11am each day. We are ushered into a compelling virtual presentation, where Abramović inducts us into being a participative community. </p>
<p>She tells us performance is the most difficult of the art forms, that you need to abandon time and surrender to the moment. Then she runs the audience through a series of Tibetan breathing exercises to make us attuned to reading the mysteries and personal language of performance artists.</p>
<h2>Durational performance</h2>
<p>Mike Parr’s Portrait of Marina Abramović is the most extreme. A blind painting event, his eyes remain closed for the entire 12 hours. His aim was to paint four black squares, one on each side of a constructed white cube gallery space, in homage to Russian constructivist painter <a href="https://www.tate.org.uk/art/artists/kazimir-malevich-1561/five-ways-look-malevichs-black-square">Kazimir Malevich’s 1915 Black Square</a>.</p>
<p>Like Malevich, Parr says his blind painting is the creation of nothingness with a view to a rebirth. But he departs from Malevich: Parr is currently driven by the reality of the shocking events in Gaza, as set out in the painted text which starts on the walls of the show: “free Palestine” and “Gaza is a Warsaw ghetto”. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/581122/original/file-20240311-28-59z6cf.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="You can just see the word 'Gaza' from behind red paint." src="https://images.theconversation.com/files/581122/original/file-20240311-28-59z6cf.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/581122/original/file-20240311-28-59z6cf.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/581122/original/file-20240311-28-59z6cf.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/581122/original/file-20240311-28-59z6cf.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/581122/original/file-20240311-28-59z6cf.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/581122/original/file-20240311-28-59z6cf.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/581122/original/file-20240311-28-59z6cf.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">Mike Parr’s work started with words looking at the war in Gaza.</span>
<span class="attribution"><span class="source">Andrew Beveridge/Adelaide Festival</span></span>
</figcaption>
</figure>
<p>These sentiments are amplified in his “vision” statement distributed at the performance. His impassioned text says: </p>
<blockquote>
<p>the Jewish diaspora rise up to join hands, to relinquish the obscene policies of its political leadership […] to demand justice, freedom, prosperity for the Palestinian people and an end to the oppression and antisemitism of the apartheid.</p>
</blockquote>
<p>Over the next 12 hours, Parr paints four black squares – at times from perilously high up on a ladder – to be covered with red paint in homage to Abramović’s former Yugoslavian communist background, then covered again with black. The painted squares, complete with drips of red paint running down to the floor, remained after the performance for viewers to ponder their meaning, along with a video of the entire event.</p>
<h2>Place and Country</h2>
<p>Less sensational but equally demanding was the durational performance by Collective Absentia, a Bangkok-based group in a work entitled Our Glorious Past, Our Glorious Present, Our Glorious Future: Our Glorious Spring. </p>
<figure class="align-right zoomable">
<a href="https://images.theconversation.com/files/581126/original/file-20240311-24-zwa8vu.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="A man sits with a covered head." src="https://images.theconversation.com/files/581126/original/file-20240311-24-zwa8vu.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=237&fit=clip" srcset="https://images.theconversation.com/files/581126/original/file-20240311-24-zwa8vu.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=900&fit=crop&dpr=1 600w, https://images.theconversation.com/files/581126/original/file-20240311-24-zwa8vu.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=900&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/581126/original/file-20240311-24-zwa8vu.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=900&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/581126/original/file-20240311-24-zwa8vu.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=1131&fit=crop&dpr=1 754w, https://images.theconversation.com/files/581126/original/file-20240311-24-zwa8vu.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=1131&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/581126/original/file-20240311-24-zwa8vu.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=1131&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">One performer meditated on non-violent forms of resistance to ongoing political events in Myanmar.</span>
<span class="attribution"><span class="source">Andrew Beveridge/Adelaide Festival</span></span>
</figcaption>
</figure>
<p>A member of their collective sat with his head covered and immobile in the middle of a passage-way, meditating on non-violent forms of resistance to ongoing political events in Myanmar. </p>
<p>All attendees at the event had to walk past and around this performance. Most stopped and connected with the sentiment of non-violent forms of resistance. One person even sat directly opposite the performer and meditated.</p>
<p>Christian Thomson’s postcolonial performance, Wait in Gold, involved him slowly and methodically pinning gold painted native daisies to every item of his exterior clothing so that he transforms from human into a larger flower form connected to Country. In this moving performance, he is responding to the denial of a voice as a result of the 2023 referendum outcome, and seeking refuge in the safety of Country.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/581124/original/file-20240311-20-59z6cf.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="A man covered in gold flowers." src="https://images.theconversation.com/files/581124/original/file-20240311-20-59z6cf.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/581124/original/file-20240311-20-59z6cf.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/581124/original/file-20240311-20-59z6cf.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/581124/original/file-20240311-20-59z6cf.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/581124/original/file-20240311-20-59z6cf.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/581124/original/file-20240311-20-59z6cf.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/581124/original/file-20240311-20-59z6cf.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">Christian Thomson seeks refuge in the safety of Country.</span>
<span class="attribution"><span class="source">Andrew Beveridge/Adelaide Festival</span></span>
</figcaption>
</figure>
<p>In Indonesian performance artist Melati Suryodarmo’s absorbing durational work, Amnesia, she slowly covers a large black board with a set of chalk markings. At each mark made, she utters “I’m sorry”. </p>
<p>The mark making is interspersed with her taking off her black shirt, placing it with other discarded shirts, and sewing a new one to put on. At other times she abandons mark making and moves across the floor, writhing as if in deep remorse, again uttering “I’m sorry”. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/581123/original/file-20240311-18-mi9d2h.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="A woman draws counting marks on a wall." src="https://images.theconversation.com/files/581123/original/file-20240311-18-mi9d2h.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/581123/original/file-20240311-18-mi9d2h.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/581123/original/file-20240311-18-mi9d2h.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/581123/original/file-20240311-18-mi9d2h.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/581123/original/file-20240311-18-mi9d2h.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/581123/original/file-20240311-18-mi9d2h.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/581123/original/file-20240311-18-mi9d2h.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">Indonesian performance artist Melati Suryodarmo’s absorbing durational work, Amnesia.</span>
<span class="attribution"><span class="source">Andrew Beveridge/Adelaide Festival</span></span>
</figcaption>
</figure>
<p>The promotional material accompanying her performance points to the work as an inner exploration of “untold narratives and forgotten realities of the past”. Her felt emotion in the performance is deeply persuasive, but I kept wondering about the amnesia from which Suryodarmo is recoiling: is it a deeply personal journey, or more?</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/marina-abramovic-retrospective-celebrates-the-grand-dame-of-performance-art-but-questions-the-genres-future-214415">Marina Abramović retrospective celebrates the grand dame of performance art – but questions the genre's future</a>
</strong>
</em>
</p>
<hr>
<h2>Encounters with AI</h2>
<p>In a different vein, Laurie Anderson’s exhibition I’ll be your mirror is an encounter with AI. Taking phrases from her song O Superman and her late husband Lou Reed’s song I’ll be your mirror, Anderson has generated intriguing text which hangs in five panels in the Adelaide Circulating Library, the city’s original lending library.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/581121/original/file-20240311-20-vh3bcl.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="A large text and two portrait photographs inside a library." src="https://images.theconversation.com/files/581121/original/file-20240311-20-vh3bcl.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/581121/original/file-20240311-20-vh3bcl.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/581121/original/file-20240311-20-vh3bcl.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/581121/original/file-20240311-20-vh3bcl.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/581121/original/file-20240311-20-vh3bcl.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/581121/original/file-20240311-20-vh3bcl.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/581121/original/file-20240311-20-vh3bcl.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">I’ll be your mirror uses AI building off songs from Laurie Anderson and Lou Reed.</span>
<span class="attribution"><span class="source">Roy VanDerVegt/Adelaide Festival</span></span>
</figcaption>
</figure>
<p>The AI generated conversations between Anderson and Reed, who passed away in 2013, oscillate between the surreal and the eerie with phrases such as</p>
<blockquote>
<p>There’s a mirror in the room <br>
And when I look at night<br>
It reflects nothing back to me. <br></p>
</blockquote>
<p>The Bible is on display and open at <a href="https://www.biblegateway.com/passage/?search=Psalm%2084-88&version=NIV">Psalms 84-88</a>, but hanging above the Bible is AI generated text based on biblical phrases, displayed as Genesis 1: 26-31. </p>
<p>A section from that text reads: </p>
<blockquote>
<p>Some nights now Noah dreams he sees his boat leave the dock<br>
It’s just another day on planet Earth <br>
Only this time it’s with an animal friend. <br></p>
</blockquote>
<p>As an adjunct to the exhibition of 21st century textual artefacts set amid 19th texts, Anderson held a virtual public conversation with the machine generating gurus she worked in Adelaide – the takeaway message being what machines generate depends on the input. </p>
<p>The exhibition is utterly intriguing, but novice viewers need an introduction to what they are about to encounter.</p>
<p><em>I’ll Be Your Mirror is on display until March 17.</em></p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/from-duchamp-to-ai-the-transformation-of-authorship-in-art-210059">From Duchamp to AI: the transformation of authorship in art</a>
</strong>
</em>
</p>
<hr>
<img src="https://counter.theconversation.com/content/223455/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Catherine Speck, with Joanna Mendelssohn, Catherine De Lorenzo and Alison Inglis, has received funding from the ARC to investigate Australian art exhibitions. </span></em></p>Ruth Mackenzie’s Adelaide Festival of Arts has two heavyweights, performance artist Marina Abramović and avant-garde artist and musician Laurie Anderson.Catherine Speck, Emerita Professor, Art History and Curatorship, University of AdelaideLicensed as Creative Commons – attribution, no derivatives.