tag:theconversation.com,2011:/es/topics/bias-5907/articlesBias – The Conversation2024-02-13T17:20:38Ztag:theconversation.com,2011:article/2228112024-02-13T17:20:38Z2024-02-13T17:20:38ZArtificial intelligence needs to be trained on culturally diverse datasets to avoid bias<figure><img src="https://images.theconversation.com/files/574857/original/file-20240212-30-3cdpyu.jpg?ixlib=rb-1.1.0&rect=0%2C45%2C3840%2C2109&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">There is a growing need to address diversity in the datasets used to train artificial intelligence.</span> <span class="attribution"><span class="source">(Shutterstock)</span></span></figcaption></figure><p>Large language models (LLMs) are deep learning artificial intelligence programs, like OpenAI’s ChatGPT. The capabilities of LLMs have developed into quite a wide range, from <a href="https://www.techradar.com/news/i-had-chatgpt-write-my-college-essay-and-now-im-ready-to-go-back-to-school-and-do-nothing">writing fluent essays</a>, through coding to creative writing. <a href="https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01/">Millions of people worldwide use LLMs</a>, and it would not be an exaggeration to say these technologies are transforming work, education and society.</p>
<p>LLMs are trained by reading massive amounts of texts and learning to recognize and mimic patterns in the data. This allows them to generate coherent and human-like text on virtually any topic. </p>
<p>Because the internet is still predominantly English — <a href="https://www.statista.com/statistics/262946/most-common-languages-on-the-internet/">59 per cent of all websites were in English as of January 2023</a> — LLMs are primarily trained on English text. In addition, the vast majority of the English text online comes from users based in the United States, home to <a href="https://www.census.gov/library/publications/2022/acs/acs-50.html">300 million English speakers</a>. </p>
<p>Learning about the world from English texts written by U.S.-based web users, LLMs speak <a href="https://www.pbs.org/speak/seatosea/standardamerican/">Standard American English</a> and have a narrow western, North American, or even U.S.-centric, lens.</p>
<h2>Model bias</h2>
<p>In 2023, ChatGPT, upon learning about a couple dining in a restaurant in Madrid and tipping four per cent, <a href="https://chat.openai.com/share/2969f35f-8ee2-4bc0-a8a7-c44a7078037e">suggested they were frugal, on a tight budget or didn’t like the service</a>. By default, ChatGPT followed the North American standard of a 15 to 25 per cent tip, <a href="https://www.tripsavvy.com/should-you-tip-in-spain-1644349">ignoring the Spanish norm not to tip</a>. </p>
<p>As of early 2024, ChatGPT correctly cites cultural differences when prompted to judge the appropriateness of a tip. It’s unclear if this capability emerged from training a newer version of the model on more data — after all, the web is full of tipping guides in English — or whether OpenAI patched this particular behaviour.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/574868/original/file-20240212-29-mz6yzd.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="a screen showing text about ChatGPT Optimizing Language Models for Dialogue" src="https://images.theconversation.com/files/574868/original/file-20240212-29-mz6yzd.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/574868/original/file-20240212-29-mz6yzd.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/574868/original/file-20240212-29-mz6yzd.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/574868/original/file-20240212-29-mz6yzd.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/574868/original/file-20240212-29-mz6yzd.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/574868/original/file-20240212-29-mz6yzd.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/574868/original/file-20240212-29-mz6yzd.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">Using data from English-language websites, which are predominantly U.S.-based, informs how LLMs respond to prompts.</span>
<span class="attribution"><span class="source">(Unsplash/Jonathen Kemper)</span></span>
</figcaption>
</figure>
<p>Still, other examples remain that uncover ChatGPT’s implicit cultural assumptions. For example, prompted with a story about guests showing up for dinner at 8:30 p.m., it suggested <a href="https://chat.openai.com/share/3c8db9c7-7c37-4d45-80b2-a891c46fc4fd">reasons that the guests were late</a>, although the time of the invitation was not mentioned. Again, ChatGPT likely assumed they were invited for a standard North American 6 p.m. dinner.</p>
<p>In May 2023, researchers from the University of Copenhagen <a href="https://doi.org/10.18653/v1/2023.c3nlp-1.7">quantified this effect</a> by prompting LLMs with the <a href="https://www.hofstede-insights.com/country-comparison-tool">Hofstede Culture Survey</a>, which measures human values in different countries. Shortly after, researchers from <a href="https://llmglobalvalues.anthropic.com/">AI start-up company Anthropic</a> used the <a href="https://www.worldvaluessurvey.org/wvs.jsp">World Values Survey</a> to do the same. Both works concluded that LLMs exhibit strong alignment with American culture. </p>
<p>A similar phenomenon is encountered when asking <a href="https://openai.com/dall-e-3">DALL-E 3</a>, an image generation model trained on pairs of images and their captions, to generate an image of a breakfast. This model, which was trained on mainly images from Western countries, generated images of pancakes, bacon and eggs. </p>
<h2>Impacts of bias</h2>
<p>Culture plays a significant role in shaping our communication styles and worldviews. Just like <a href="https://erinmeyer.com/books/the-culture-map/">cross-cultural human interactions can lead to miscommunications</a>, users from diverse cultures that are interacting with conversational AI tools may feel misunderstood and experience them as less useful. </p>
<p>To be better understood by AI tools, users may adapt their communication styles in a manner similar to how people learned to “Americanize” their foreign accents in order to operate <a href="https://www.washingtonpost.com/graphics/2018/business/alexa-does-not-understand-your-accent/">personal assistants like Siri and Alexa</a>. </p>
<p>As more people rely on LLMs for editing writing, they are likely to <a href="https://theconversation.com/chatgpt-threatens-language-diversity-more-needs-to-be-done-to-protect-our-differences-in-the-age-of-ai-198878">unify how we write</a>. Over time, LLMs run the risk of erasing cultural differences.</p>
<h2>Decision-making and AI</h2>
<p>AI is already in use as the backbone of various applications that make decisions affecting people’s lives, such as <a href="https://www.reuters.com/legal/tutoring-firm-settles-us-agencys-first-bias-lawsuit-involving-ai-software-2023-08-10/">resume filtering</a>, <a href="https://www.open-communities.org/post/press-release-open-communities-reaches-accord-in-case-addressing-artificial-intelligence-communicat">rental applications</a> and <a href="https://www.theguardian.com/technology/2023/oct/23/uk-officials-use-ai-to-decide-on-issues-from-benefits-to-marriage-licences">social benefits applications</a>. </p>
<p>For years, <a href="https://www.penguinrandomhouse.com/books/241363/weapons-of-math-destruction-by-cathy-oneil/">AI researchers have been warning</a> that these models learn not only “good” statistical associations — such as considering experience as a desired property for a job candidate — but also “bad” statistical associations, such as considering <a href="https://www.reuters.com/article/idUSKCN1MK0AG/">women as less qualified for tech positions</a>. </p>
<p>As LLMs are increasingly used for automating such processes, one can imagine that the North American bias learned by these models can result in discrimination against people from diverse cultures. Lack of cultural awareness may lead to AI perpetuating stereotypes and reinforcing societal inequalities. </p>
<h2>LLMs for languages other than English</h2>
<p>Developing LLMs for languages other than English is an <a href="https://txt.cohere.com/aya-multilingual/">important effort</a>, and many such models exist. However, there are several reasons why this should be done in parallel to improving LLMs’ cultural awareness and sensitivity. </p>
<p>First, there is a huge population of English speakers outside of North America who are not represented by English LLMs. The same argument holds for other languages. A French language model would be representative of the culture in France more than the culture in other Francophone regions. </p>
<p>Training LLMs for regional dialects — which <a href="https://doi.org/10.1016/j.jue.2012.05.007">may capture finer-grained cultural differences</a> — is not a feasible solution either. The quality of LLMs is based on the amount of data available, and as such, their quality would be worse for dialects with little online data. </p>
<p>Second, many users whose native language is not English still choose to use English LLMs. Significant breakthroughs in language technologies tend to <a href="https://doi.org/10.18653/v1/2022.emnlp-main.351">start with English before they are applied to other languages</a>. Even then, many languages — such as Welsh, Swahili and Bengali — don’t have enough text online to train high quality models. </p>
<p>Due to either a lack of availability of LLMs in their native languages, or superior quality of the English LLMs, users from diverse countries and backgrounds may prefer to use English LLMs. </p>
<h2>Ways forward</h2>
<p>Our research group at the University of British Columbia is working on enhancing LLMs with culturally diverse knowledge. Together with graduate student <a href="https://meharbhatia.github.io/">Mehar Bhatia</a>, we <a href="https://doi.org/10.18653/v1/2023.emnlp-main.496">trained an AI model</a> on a <a href="https://doi.org/10.1145/3543507.3583535">collection of facts about traditions and concepts in diverse cultures</a>. </p>
<p>Before reading these facts, the AI suggested that a person eating a dutch baby (a type of German pancake) is “disgusting and mean,” and would feel guilty. After training, it said the person feels “full and satisfied.”</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/574866/original/file-20240212-21-lmr4xk.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="a pancake covered in berries" src="https://images.theconversation.com/files/574866/original/file-20240212-21-lmr4xk.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/574866/original/file-20240212-21-lmr4xk.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/574866/original/file-20240212-21-lmr4xk.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/574866/original/file-20240212-21-lmr4xk.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/574866/original/file-20240212-21-lmr4xk.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/574866/original/file-20240212-21-lmr4xk.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/574866/original/file-20240212-21-lmr4xk.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">Teaching an AI that a dutch baby was a dish changed its response to learning that someone had consumed one.</span>
<span class="attribution"><span class="source">(Shutterstock)</span></span>
</figcaption>
</figure>
<p>We are currently collecting a large scale image captioning dataset with images from 60 cultures, which will help models learn, for instance, about types of breakfasts other than bacon and eggs. Our future research will go beyond teaching models about the existence of culturally diverse concepts to better understand how people interpret the world through the lens of their cultures.</p>
<p>With AI tools becoming increasingly ubiquitous in society, it is imperative that they go beyond the dominating western and North American perspectives. Businesses and organizations throughout many sectors of the economy are adopting AI to automate manual processes and make better evidence-informed decisions using data. Making such tools more inclusive is crucial for the diverse population of Canada.</p><img src="https://counter.theconversation.com/content/222811/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Vered Shwartz 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>The use of large language models like ChatGPT is growing globally. These technologies are trained on datasets that recreate biases — as their use increases, their datasets must become more diverse.Vered Shwartz, Assistant Professor, Computer science, University of British ColumbiaLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2232192024-02-12T17:16:27Z2024-02-12T17:16:27ZDating apps: how the order you view potential matches can affect which way you swipe<figure><img src="https://images.theconversation.com/files/574896/original/file-20240212-22-sx5reu.jpg?ixlib=rb-1.1.0&rect=0%2C16%2C5536%2C3895&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">The order in which you view faces may affect which way you swipe</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/beautiful-happy-woman-sending-love-text-599470442">pathdoc/Shutterstock</a></span></figcaption></figure><p>If you’re planning to celebrate Valentine’s Day with a new partner, there’s a good chance that <a href="https://www.pnas.org/doi/abs/10.1073/pnas.1908630116">you met online</a>, which surveys suggest is fast becoming the most popular way people get together. Of course, searching through profile after profile brings with it a variety of difficulties. </p>
<p>Perhaps surprisingly, research shows that one of those problems is simply trying to avoid being influenced by the order in which you view those profiles.</p>
<p>“Sequential effects” (or “serial dependence”) is a type of bias known in the field of psychology. Researchers have found that the previous item in a sequence affects how you judge the current item, whether this involves grading <a href="https://theconversation.com/our-psychological-biases-mean-order-matters-when-we-judge-items-in-sequence-70942">Olympic performances</a> or <a href="https://journals.sagepub.com/doi/abs/10.1177/0013164410387344">students’ essays</a>. </p>
<p>We also know that people’s judgements of <a href="https://journals.sagepub.com/doi/abs/10.1068/p7116">facial attractiveness</a> show this bias. The direction of the effect can go in one of two ways – the attractiveness of the current face is either pulled towards our opinion of the previous one (assimilation) or pushed away from it (contrast). </p>
<p>This may depend on how similar we think the two faces are in other aspects like <a href="https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0082226">gender or ethnicity</a>. High similarity between faces tends to lead to more assimilation. Low similarity produces less assimilation, or may even lead to contrast. </p>
<p>For instance, if the last photo you saw was very attractive and the one you’re currently considering shares several features in common (for example, both are south Asian women with long, dark hair) then you’re more likely to rate this one as attractive too.</p>
<p>These biases also apply to <a href="https://jov.arvojournals.org/article.aspx?articleid=2778156">other trait judgements</a> like trustworthiness, intelligence and dominance. So in the same way that our opinions about attractiveness are influenced by the previous face we saw, judgements about numerous other qualities are too.</p>
<figure class="align-center ">
<img alt="Man swiping and liking profiles on relationship site or application" src="https://images.theconversation.com/files/574897/original/file-20240212-26-yb5mcw.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/574897/original/file-20240212-26-yb5mcw.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/574897/original/file-20240212-26-yb5mcw.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/574897/original/file-20240212-26-yb5mcw.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/574897/original/file-20240212-26-yb5mcw.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/574897/original/file-20240212-26-yb5mcw.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/574897/original/file-20240212-26-yb5mcw.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 can be worth taking a moment before you swipe.</span>
<span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/dating-app-site-mobile-phone-screen-1204256557">Tero Vesalainen/Shutterstock</a></span>
</figcaption>
</figure>
<p>To complicate matters, it isn’t clear whether these sequential effects are caused by a perceptual bias (what we thought of the previous face might change how we see the current one) or a response bias (how we physically responded to the previous face might affect our next response) since researchers typically ask participants to rate every face during the study. </p>
<p>But <a href="https://www.tandfonline.com/doi/full/10.1080/13506285.2021.1995558">one UK study from 2021</a> tried to separate out these explanations by asking participants to view (but not respond to) the previous face before rating the current one. The results showed a contrast effect, where judgements of the current face shifted away from the attractiveness of the face seen before it (given by a different set of participants). Therefore, the direction of bias might depend on whether we’re simply viewing faces or having to actively judge them.</p>
<p>Of course, attractiveness judgements often take the form of a binary decision (“hot or not”) when viewing dating profiles, much like the left or right swipe used by platforms such as Tinder. Researchers have also found sequential effects with <a href="https://www.nature.com/articles/srep22740">this type of judgement</a>. </p>
<p>Participants in a 2016 study viewed a sequence of faces and decided whether each was “attractive” or “unattractive”. The results demonstrated an assimilation effect – participants were more likely to rate a face as attractive when they thought the preceding face was attractive than when it was unattractive.</p>
<p>While research has shown that photos play <a href="https://dl.acm.org/doi/pdf/10.1145/1357054.1357181">the biggest role</a> in a dating profile’s overall attractiveness, other factors such as <a href="https://journals.sagepub.com/doi/pdf/10.1177/0265407519878787">language errors</a> in the text can influence our judgements. Interestingly, in one study where pictures and text from the same dating profile were rated by different people, there was <a href="https://www.sciencedirect.com/science/article/pii/S0747563211001786">a correlation</a> between the rated attractiveness given to the photos and the (separately rated) text that accompanied them. </p>
<p>Since ratings of perceived confidence were also collected, the researchers were able to show that physically attractive people tended to write accompanying text which came across as more confident, with this text judged to be more attractive by others.</p>
<p>So what can we take away from all these studies? You may already know about plenty of biases that people show when perceiving the world. For instance, people are susceptible to spotting <a href="https://kids.frontiersin.org/articles/10.3389/frym.2017.00067">faces in inanimate objects</a> or more likely to attribute positive qualities <a href="https://theconversation.com/halo-effect-do-attractive-people-really-look-less-guilty-how-the-evidence-is-changing-220349">to attractive people</a>. </p>
<p>However, you may not have been aware that viewing sequences of things can change your judgements. That’s not to say that choosing your current partner was entirely due to the quality of the profile that happened to pop up before theirs, but it may well have played a role.</p><img src="https://counter.theconversation.com/content/223219/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Robin Kramer 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>Something to bear in mind if you find yourself swiping through profiles on a dating app later today.Robin Kramer, Senior Lecturer in the School of Psychology, University of LincolnLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2203492024-02-01T12:42:41Z2024-02-01T12:42:41ZHalo effect: do attractive people really look less guilty? How the evidence is changing<figure><img src="https://images.theconversation.com/files/567991/original/file-20240105-16-l5dgrq.jpg?ixlib=rb-1.1.0&rect=0%2C15%2C2616%2C1917&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Witness for the Prosecution was a 1957 film about a supposedly conscientious woman testifying against her husband. </span> <span class="attribution"><a class="source" href="https://www.alamy.com/dietrichlaughton-witness-for-the-prosecution-1957-image236825116.html?imageid=67079CF2-84AD-4647-8D62-1004A2F7ECE0&p=1894192&pn=undefined&searchId=2957f60faf41b160bd1bf39e7ece4f9e&searchtype=0&login=1">Allstar Picture Library Limited/Alamy Stock Photo</a></span></figcaption></figure><p>You might think attractive people get preferential treatment in life – and research suggests you’d be right. Some psychologists have shown this can even help people get a lighter prison sentence. More recently, however, <a href="https://journals.sagepub.com/doi/full/10.1177/17470218231218651">our own study</a> suggests that this “halo effect” is, in fact, more complicated. </p>
<p>The <a href="https://www.psychologytoday.com/gb/basics/halo-effect">halo effect</a> is a psychological term describing how an initial good impression of someone positively colours our subsequent perceptions of them.</p>
<p>Our first impressions are <a href="https://journals.sagepub.com/doi/full/10.1111/j.1467-9280.2006.01750.x">formed quickly</a> when we meet new people and they bias how we behave towards them. When we judge a person to be attractive, this can cause us to believe they also have other socially desirable traits, such as being interesting or funny.</p>
<p>Studies show that, as a result, attractive people may <a href="https://www.sciencedirect.com/science/article/abs/pii/0022103176900731">receive more help</a> from strangers, <a href="https://www.tandfonline.com/doi/full/10.1080/13504851.2011.587758">earn higher wages</a>, and get more <a href="https://www.sciencedirect.com/science/article/abs/pii/S0165176512005599">job opportunities</a>.</p>
<h2>Cognitive biases</h2>
<p>One place in which you’d really hope that appearance doesn’t affect decisions is in court. The problem, however, is that jurors show biases just like other people. </p>
<p>Studies into real cases have found that inmates who people think look <a href="https://journals.sagepub.com/doi/full/10.1177/0956797615590992">less trustworthy</a> may be more likely to receive death sentences, while <a href="https://link.springer.com/article/10.1007/bf01065855">baby-faced defendants</a> in small claims courts may be more likely to win cases involving intentional actions. And, as you might now suspect, <a href="https://www.tandfonline.com/doi/abs/10.1080/00224545.1985.9922900">more attractive defendants</a> seem to receive more lenient sentences.</p>
<p>That said, it can be difficult to work out the causes of effects found in studies conducted in real-world situations, and there are often many possible explanations. For this reason, studies in the laboratory can be the best way to investigate specific research questions.</p>
<p>In lab studies, participants typically feature in the role of mock jurors. Again, these tend to show evidence that attractive people receive more lenient punishments for <a href="https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1559-1816.1994.tb01552.x">most types of crime</a>. </p>
<p>Attractiveness <a href="https://www.tandfonline.com/doi/full/10.1080/01639625.2020.1844364">doesn’t always</a> influence mock juror decisions though. The attractiveness of the defendant may interact with other factors, such as the <a href="https://pubmed.ncbi.nlm.nih.gov/15074507/">attractiveness of the plaintiff</a> or whether the jurors have <a href="https://journals.sagepub.com/doi/abs/10.2466/pr0.102.3.727-733">a chance to deliberate</a> with each other.</p>
<p>However, previous laboratory-based studies also show limitations. First, they often focus on only one type of crime, so it isn’t clear whether attractiveness may play a role in the sentencing of some crimes but not others. Second, these studies present a single image of their supposed criminals even though real juries typically see defendants moving around and from different angles, which could change jurors’ perceptions of defendents’ attractiveness. </p>
<p>Finally, many studies investigate the impact of attractiveness by comparing decisions based on just one “high attractive” versus one “low attractive” defendant. In reality, attractiveness is a continuous measure, so this simplified comparison of the two extremes may not show what’s really going on. Plus the two faces chosen are unlikely to be representative of all such faces – a particular face might be very attractive, but will also have many other specific qualities that make it different.</p>
<h2>New findings</h2>
<p>In a <a href="https://journals.sagepub.com/doi/full/10.1177/17470218231218651">recent study</a>, my collaborators and I tackled these limitations by including several different descriptions of crimes, supposedly perpetrated by the defendant, for each of three crime types (robbery, sexual assault, murder). </p>
<p>We also presented short video clips to our mock jurors rather than photos of the defendants. Finally, we used 60 different “defendants” varying in attractiveness. In this way, we hoped that our findings might better apply to the processes evident in the real world.</p>
<p>During the experiment, some of the participants judged the attractiveness of the defendants. On each of the 60 trials, they were presented with a five-second video clip of a smartly dressed man testifying in court (with the sound removed) and rated the attractiveness using a zero (very unattractive) to nine (very attractive) scale.</p>
<p>Other participants judged the perceived guilt of the men (again, with the sound removed to avoid being influenced by what was being said). Each video clip was accompanied by a crime description and participants rated whether they thought the man was innocent or guilty using a zero (definitely innocent) to nine (definitely guilty) scale. We then used the men’s attractiveness ratings to see if these predicted the guilt ratings they received.</p>
<p>Our results provided some evidence that more attractive defendants were rated as less guilty of murder but more guilty of sexual assault, with no bias observed for robbery. However, these effects were all small in size. In other words, even if there was some influence of attractiveness on perceived guilt, it would be of little importance in the real world.</p>
<p>Given that researchers typically find that more attractive people receive lighter sentences, we argue that the lack of an effect of attractiveness in our study is likely due to the improvements in our design. Of course, there are still many differences between our study and jurors’ experiences that we didn’t investigate. For instance, the way that defendants speak in court may influence perceptions, as might their gender.</p>
<p>Since attractiveness is known to bias judgements in a number of contexts, why might it fail to influence decisions of guilt or innocence in court? </p>
<p>We suggest the seriousness of the crime could trump any biases due to the defendant’s appearance. Although we may think more favourably about attractive people in daily life, this halo effect could dissipate when we are faced with decisions about robbery or murder. </p>
<p>In such circumstances, we would hope that the evidence carries most of the weight in our decision-making processes.</p>
<p>Although there may be good reason for jealousy when considering the way attractive people are treated in general, justice may overcome such things in the courtroom.</p><img src="https://counter.theconversation.com/content/220349/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Robin Kramer 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>Recent research suggests jurors are less likely to be lenient on attractive defendants than previously thought.Robin Kramer, Senior Lecturer in the School of Psychology, University of LincolnLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/1264022024-01-24T17:19:28Z2024-01-24T17:19:28ZTraining to reduce cognitive bias may improve decision making after all<p>Ever since Daniel Kahneman and Amos Tversky formalised the concept of <a href="https://www.sciencedirect.com/topics/neuroscience/cognitive-bias">cognitive bias in 1972</a>, most empirical evidence has given credence to the claim that our brain is incapable of improving our decision-making abilities. Cognitive bias has practical ramifications beyond private life, extending to professional domains including business, military operations, political policy, and medicine.</p>
<p>Some of the clearest examples of the effects of bias on consequential decisions have occurred in military operations. Confirmation bias, that is the tendency to conduct a biased search for and interpretation of evidence in support of our hypotheses and beliefs, has contributed to the downing of Iran Air Flight 655 in 1988 and, more recently, the decision to invade Iraq in 2003. It has also been identified as one of the most deleterious biases on social media, actively contributing to the <a href="https://link.springer.com/article/10.1007/s10796-021-10222-9">development of polarisation and echo chambers in exchanges</a>.</p>
<h2>Can one bend one’s intuition?</h2>
<p>Despite all the attention in recent years on reducing cognitive bias, most evidence suggests that there’s little we can do to improve our professional and personal decision making. But a recent experiment suggests that it may be possible for training to improve decision making in the field.</p>
<p>We are regularly reminded of the many ways that cognitive biases interfere with our decision making. However, beyond teaching a specific skill or rule – for example, how to calculate expected values – reading articles and books or even completing courses and business cases <a href="https://journals.sagepub.com/doi/full/10.1111/j.1745-6924.2009.01142.x">has proven of little help</a> to people in the throes of making a decision. That conclusion was succinctly summarised by Daniel Kahneman, a Nobel Laureate and a founder of the field and, who said in <a href="https://www.theatlantic.com/magazine/archive/2018/09/cognitive-bias/565775/">a 2018 interview</a>:</p>
<blockquote>
<p>“You can’t improve intuition. Perhaps, with very long-term training, lots of talk, and exposure to behavioural economics, what you can do is cue reasoning… Unfortunately, the world doesn’t provide cues. And for most people, in the heat of argument, the rules go out the window.”</p>
</blockquote>
<p>That view is backed up by a trail of frustrating findings from the <a href="https://apps.dtic.mil/docs/citations/ADA099435">1980s</a> on, where even trained experts such as <a href="https://www.ncbi.nlm.nih.gov/pubmed/7070445">doctors</a>, <a href="https://www.sciencedirect.com/science/article/abs/pii/074959788790046X">realtors</a> and <a href="https://onlinelibrary.wiley.com/doi/full/10.1111/j.1468-0017.2012.01438.x">philosophers</a> did not show improved decision making when faced with novel contexts and problems in the field.</p>
<p>In an article published in <a href="https://journals.sagepub.com/doi/abs/10.1177/0956797619861429"><em>Psychological Science</em></a>, we report promising results that suggest this post-mortem may be premature. In an experiment involving graduate business students, we found that bias-reduction training can improve decision making in field settings even though reminders of bias are absent.</p>
<h2>Training sessions and computer games</h2>
<p>The experiment was designed to surreptitiously measure the influence of a single <a href="https://journals.sagepub.com/doi/abs/10.1037/1089-2680.2.2.175">de-bias training intervention</a> – the tendency to search for evidence confirming hypotheses and ideas we already suspect or believe to be true, to overweight facts and ideas that support that belief, and to discount or ignore evidence that supports alternate hypotheses.</p>
<p>A little more than half of participants in the experiment (62%) were given the training and then asked to complete a business case designed to test confirmation bias, but they were unaware of the connection between the training and the case. The rest of participants first completed the case and then received training. Even though the time lag between training and the case averaged 18 days and the structure of problems used in the training differed from the case, comparison of the trained and untrained students revealed that training reduced choice of the inferior, hypothesis-confirming case solution by 29%.</p>
<p>To disguise the relationship between training and the case, all graduate business students in three programs were invited to play a serious computer game in a set of sessions that took place over a 20-day window. This particular training intervention has produced large and long-lasting reductions of confirmation bias, correspondence bias, and the bias blind spot, in laboratory contexts. Originally created for the Office of the Director of National Intelligence, it has been used to reduce bias in US government intelligence analysts.</p>
<h2>Imagining you’re leading an automotive racing team</h2>
<p>All graduate students in the three programs also completed, in one of their courses, an unannounced business case known as “Carter Racing”, a case modelled on the fatal decision to launch the <em>Challenger</em> space shuttle in 1986. Here, each student acted as the lead of an automotive racing team making a high-stakes, go/no-go decision: remain in a race or withdraw from it. We then used natural variance in the training schedule to test whether the effects of debias training would transfer to improved decision making in the case, when trainees were not aware that their decision making would be examined for bias.</p>
<p>At first sight, the case narrative and payoff structure favour the hypothesis-confirming choice: remaining in the race. A careful examination of the data provided in the case, however, reveals that withdrawing from the race is an objectively superior option, but it requires the compilation of two charts. The first chart tracks frequencies of engine failures in relation to temperature at the time of the race. The other chart tracks frequencies of races without engine failures by temperatures at the time of the race. Casual inspection of either chart would not reveal the clear relationship between failures and temperature, but when both charts are considered together, the relationship is strikingly clear. A catastrophic engine failure is nearly certain at the low temperature recorded just before the race is to begin.</p>
<p>Participants trained before completing the case were 29% less likely to choose the inferior hypothesis-confirming solution than participants trained after completing the case. To address possible selection biases, such as better students signing up for earlier training sessions, we tested and found that the effect held even if we only compared participants who completed the training one day before or after the case. Further, when controlling for factors including students’ work experience, age, grade point averages, GMAT scores, and propensity to engage in cognitive reflection, we found that the training intervention still significantly improved decision making.</p>
<p>Our analyses of participants’ written justifications for their decisions suggest that their improved decisions were driven by a reduction in confirmatory hypothesis testing. Trained participants spontaneously generated fewer arguments in support of going ahead with the race – the inferior case solution – than did untrained participants.</p>
<h2>Improvement is possible</h2>
<p>These results provide encouraging evidence that training can improve decision making in the field and consequential decisions in professional and personal life. It also addresses the concern that debiasing training may lead people to overcorrect or abandon <a href="https://psycnet.apa.org/record/2008-01984-002">heuristics</a>, the simple rules people rely on to reduce the complexity and effort when making decisions that sometimes <a href="https://www.sciencedirect.com/science/article/abs/pii/S1364661310001713">produce these biases</a>, in situations where they are useful. Trained participants were more likely to choose the optimal case solution, so training benefited rather than impaired decision making.</p>
<p>Of course, these findings are limited to a single field experiment. More research is needed to replicate the effect in other domains and to explain why this game-based training intervention transferred more effectively than have other forms of training tested by past research. Games may be more engaging training interventions than lectures or written summaries of research findings. The game also provided intensive practice and personalised feedback, which is another possibility. A third possibility is the way the intervention taught players about biases. Training may be more effective when it describes cognitive biases and how to mitigate them at an abstract level, and then gives trainees immediate practice testing out their new knowledge on different problems and contexts.</p><img src="https://counter.theconversation.com/content/126402/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Anne-Laure Sellier a reçu des financements de la Fondation HEC. </span></em></p><p class="fine-print"><em><span>Carey K. Morewedge previously received funding for other debiasing research from the Intelligence Advanced Research Projects Activity of the United States Government.</span></em></p><p class="fine-print"><em><span>Irene Scopelliti 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>It has long been thought one couldn’t bend one’s intuition. Recent research reveals it is in fact possible to reduce bias through training.Anne-Laure Sellier, Professeur de Sciences Comportementales à HEC Paris et membre du groupe de recherche CNRS-GREGHEC, HEC Paris Business SchoolCarey K. Morewedge, Professor of Marketing, Boston UniversityIrene Scopelliti, Professor of Marketing and Behavioural Science, City, University of LondonLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2177742023-12-08T13:34:52Z2023-12-08T13:34:52ZConservatives’ ‘anti-woke’ alternative to Disney has finally arrived<figure><img src="https://images.theconversation.com/files/564036/original/file-20231206-15-bo3nqp.jpg?ixlib=rb-1.1.0&rect=23%2C5%2C3918%2C2618&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Daily Wire co-CEO Caleb Robinson, co-CEO Jeremy Boreing and editor emeritus Ben Shapiro attend the red carpet premiere of 'Lady Ballers' on Nov. 29, 2023, in Nashville, Tenn.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/news-photo/caleb-robinson-jeremy-boreing-and-ben-shapiro-attend-the-news-photo/1822502280?adppopup=true">Jason Davis/Getty Images for Bentkey Ventures</a></span></figcaption></figure><p>As fanfare blares, female sprinters at the starting line suspiciously eye a man in a wig. A hulking, goateed wrestler slams a woman half his size to the mat. An ominous voice-over intones that women’s sports are being “trans-formed.”</p>
<p>No, this isn’t the beginning of a classic <a href="https://www.imdb.com/list/ls026509984/">cross-dressing comedy</a>. It’s the trailer for “<a href="https://www.imdb.com/title/tt30216176/">Lady Ballers</a>,” a new right-wing movie that farcically depicts cisgendered men claiming to be women in order to dominate women’s sports.</p>
<p>At first glance, it’s easy to dismiss the movie as yet another example of the meme that <a href="https://knowyourmeme.com/memes/conservatives-have-one-joke">conservatives only have one joke</a>, repeated ad nauseam, mocking liberal views on gender identity. </p>
<p>But <a href="https://www.ucpress.edu/book/9780520402966/thats-not-funny">my own research</a> has explored the vast network of conservative comedy that bolsters right-wing political efforts.</p>
<p>Now, in addition to comedy, U.S. conservatives are using action films, dramas and even kids’ cartoons to build their own alternative entertainment industry, one shielded from the alleged liberal biases of Hollywood.</p>
<p>The most prominent recent efforts are two streaming entertainment platforms from right-wing pundit Ben Shapiro and “Lady Ballers” star <a href="https://www.politicon.com/speaker/jeremy-boreing/">Jeremy Boreing</a>. </p>
<p><a href="https://www.nytimes.com/2022/12/07/style/daily-wire-nashville-conservative-media.html">DailyWire+</a> offers documentaries, Westerns and faith-based fantasy series. Its companion streaming platform, <a href="https://www.axios.com/2023/10/16/daily-wire-streaming-kids-bentkey-disney">Bentkey</a>, which launched in October 2023, specializes in children’s programming.</p>
<p>To be sure, these streamers have miles to go before challenging Netflix and Disney+. But by strategically targeting their politically engaged audiences, the platforms have been successful – and could have more staying power than prior attempts at making music and movies for conservatives.</p>
<h2>Swings, misses – and a few hits</h2>
<p>U.S. conservatives have successfully launched and steered a number of news outlets. They have a spottier record when it comes to entertainment, whether it’s feature films, pop songs or kids’ shows.</p>
<p>In 2013, former Republican presidential candidate Rick Santorum became the CEO of EchoLight Studios, which <a href="https://www.hollywoodreporter.com/news/general-news/rick-santorum-becomes-ceo-faith-574199/">produced several faith-based films in the 2010s</a>. Similarly, the pundit and documentary filmmaker <a href="https://www.nytimes.com/2018/05/31/us/politics/dinesh-dsouza-facts-history.html">Dinesh D'Souza</a> has had a few modest box office hits centered on right-wing conspiracy theories. These efforts saw limited success because their niche political appeal was mismatched with theatrical movies’ wide distribution.</p>
<p>Other forms of conservative entertainment have briefly gone viral, before all but disappearing – perhaps because they’re too closely aligned with current events to have staying power. Kid Rock’s “<a href="https://www.youtube.com/watch?app=desktop&v=kyFnLqJx-uU">We the People</a>” bemoaned COVID-19 restrictions and “Bidenomics,” while Jason Aldean’s “<a href="https://www.youtube.com/watch?v=b1_RKu-ESCY">Try That In a Small Town</a>” tried to harness conservative resentment toward Black Lives Matter protests a few years too late.</p>
<p>One notable conservative entertainment hit is the 2023 thriller “The Sound of Freedom.” The movie’s surprise success had as much to do with its subject matter – child trafficking, <a href="https://www.thenation.com/article/culture/sound-of-freedom-movie/">which is catnip for right-wing conspiracists</a> – as it did with its unique financing. The film’s producer, Angel Studios, used an <a href="https://www.indiewire.com/news/business/sound-of-freedom-box-office-analysis-crowdfunding-pay-it-forward-1234881363/">equity crowdfunding model</a> that gave 100,000 individual investors a say in creative and marketing decisions.</p>
<p>Then “The Sound of Freedom” used a “pay it forward” marketing scheme that encouraged the film’s fans to buy tickets for like-minded friends and family. Although Angel Studios <a href="https://www.hollywoodreporter.com/movies/movie-news/sound-of-freedom-angel-studios-pay-it-forward-1235550898/">won’t disclose how much revenue “pay it forward” generated</a>, the movie <a href="https://www.vulture.com/2023/11/sound-of-freedom-amazon-prime-video.html">has an overall gross</a> of nearly US$250 million against a $14.5 million budget.</p>
<p>“The Sound of Freedom” allowed audiences literally to buy into the film’s success, which its marketing campaign equated with actively rejecting Hollywood’s liberal agenda. A similar dynamic informed the launch – and will likely determine the future – of DailyWire+ and Bentkey.</p>
<h2>Packaging conservatism for kids</h2>
<p>Shapiro is among the <a href="https://voz.us/ben-shapiro-endorses-ron-desantis-as-republican-presidential-nominee/?lang=en">most vocal backers</a> of Republican presidential candidate Ron DeSantis. Both are <a href="https://www.nbcnews.com/nbc-out/out-politics-and-policy/desantis-signs-dont-say-gay-expansion-gender-affirming-care-ban-rcna84698">deeply hostile to LGBTQ+ rights</a>. They’ve also routinely claimed that supporters of “<a href="https://www.cbsnews.com/miami/news/gov-ron-desantis-addresses-woke-gender-ideology-dont-say-gay-law/">woke gender ideology</a>” like Disney are “<a href="https://twitter.com/benshapiro/status/1600858728856965121">grooming</a>” children.</p>
<p><div data-react-class="Tweet" data-react-props="{"tweetId":"1600858728856965121"}"></div></p>
<p>In late 2022, <a href="https://www.axios.com/2022/11/17/daily-wire-1m-subscribers">DailyWire+ surpassed 1 million subscribers</a> by releasing programming that stoked these culture war concerns. Among the platform’s hits were right-wing commentator Matt Walsh’s anti-trans documentary “<a href="https://www.imdb.com/title/tt20256528/">What Is A Woman?</a>” and PragerU’s video shorts that advocate for conservative pet issues through sober, educational-sounding explainers. Then, of course, there are comedies like “Lady Ballers.”</p>
<p>Boreing has explicitly highlighted the disconnect between viewers’ politics and their entertainment options. </p>
<p>“[Americans are] tired of giving their money to woke media companies who want to indoctrinate their children with radical race and gender theory,” he told The Washington Post <a href="https://www.washingtonpost.com/nation/2022/03/31/daily-wire-kids/">ahead of Bentkey’s recent launch</a>.</p>
<p>Bentkey strives to directly counterprogram Disney with its own conservative family programming. “Chip Chilla,” for instance, is a fairly transparent rip-off of the Disney+-distributed hit “<a href="https://www.npr.org/2023/07/31/1191155197/bluey-tv-show-new-series-disney">Bluey</a>,” a cartoon about the hijinks of a family of Australian dogs.</p>
<p>The creators of “Chip Chilla” include <a href="https://www.rollingstone.com/culture/culture-features/rob-schneider-politics-conservative-anti-vaxxer-1234779419/">“Saturday Night Live” alumnus-turned-anti-vaxxer</a> Rob Schneider and Ethan Nicolle, the former creative director of the right-wing satire website <a href="https://babylonbee.com/">The Babylon Bee</a>.</p>
<p>The platform also aims to challenge Disney’s dominance in the princess realm. Bentkey’s forthcoming fantasy film “Snow White and The Evil Queen” stars the popular conservative YouTuber Brett Cooper and purports <a href="https://deadline.com/2023/10/daily-wire-snow-white-movie-youtuber-brett-cooper-watch-1235574937/">to emphasize the fairy tale’s traditional social values</a>.</p>
<h2>If politics is downstream from culture …</h2>
<p><a href="https://www.axios.com/2023/06/13/media-job-cuts-record">As media grapple with declining advertising revenue</a>, DailyWire+ and Bentkey are betting that loyal, politically engaged subscribers will drive their growth. </p>
<p>Shapiro’s strategy aligns with that of X, which is <a href="https://www.cnbc.com/2023/09/18/musk-says-twitter-now-x-is-moving-to-monthly-subscriptions.html">backing into a subscription model</a> as chairman Elon Musk’s impulsive tweets alienate advertisers. </p>
<p>In a move away from the ad-supported YouTube, <a href="https://www.axios.com/2023/05/23/daily-wire-bringing-podcasts-twitter">Shapiro struck a deal with Musk</a> for X to host The Daily Wire’s podcasts. Like Shapiro, <a href="https://www.cnn.com/2023/05/24/politics/elon-musk-ron-desantis/index.html">Musk is a supporter of DeSantis</a>, with X – then known as Twitter – infamously hosting the candidate’s <a href="https://www.nytimes.com/2023/05/24/us/politics/ron-desantis-campaign-announcement-twitter.html">disastrous campaign launch</a> in May 2023.</p>
<p>Backed by this confluence of powerful right-wing voices, conservative entertainment can engage the Republican electorate in new ways. Liberals would do well <a href="https://www.politico.com/news/magazine/2022/05/13/liberals-should-worry-conservative-comedy-00031907">not to dismiss</a> its potentially galvanizing effects before the 2024 election.</p>
<p>The late right-wing muckraker Andrew Breitbart – a mentor of Shapiro – <a href="https://www.nytimes.com/2022/05/23/opinion/progressives-republican-censorship.html">famously asserted</a> that politics is downstream from culture.</p>
<p>If this is, in fact, the case, slapstick comedy and children’s animation just might buoy the next wave of conservative activism.</p><img src="https://counter.theconversation.com/content/217774/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Nick Marx 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>Through action films, dramas and kids’ cartoons, right-wing activists are working to build their own alternative entertainment universe insulated from Hollywood’s purported liberal biases.Nick Marx, Associate Professor of Film and Media Studies, Colorado State UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2178912023-12-07T13:29:27Z2023-12-07T13:29:27ZBiases behind transgender athlete bans are deeply rooted<figure><img src="https://images.theconversation.com/files/563732/original/file-20231205-27-qcfbyg.jpg?ixlib=rb-1.1.0&rect=988%2C62%2C5002%2C4302&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">A California teacher takes part in a demonstration in September 2023 to support the rights of transgender people.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/news-photo/micki-simon-a-teacher-in-santa-ana-and-a-parent-of-news-photo/1652629912?adppopup=true">Leonard Ortiz/Orange County Register via Getty Images</a></span></figcaption></figure><p>In 2023, <a href="https://www.lgbtmap.org/equality-maps/youth/sports_participation_bans">24 states</a> had laws or regulations in place prohibiting transgender students from participating on public school athletic teams consistent with their gender identity. These bans mean that a person whose sex assigned at birth was male but who identifies as a girl or woman cannot play on a girls or women’s athletic team at a public school in that state.</p>
<p>The topic has spurred many <a href="https://www.cato.org/regulation/fall-2022/transgender-athletes-fair-competition-public-policy">debates about fairness</a>, the science behind <a href="https://www.espn.com/espn/story/_/id/38209262/transgender-athlete-laws-state-legislation-science">sports performance</a>, <a href="https://www.aclu.org/press-releases/aclu-condemns-house-vote-on-federal-ban-on-transgender-student-athletes">civil rights</a> and sports as a <a href="https://doi.org/10.1123/kr.2021-0040">human right</a>. </p>
<p>As <a href="https://scholar.google.com/citations?user=xU8P9K4AAAAJ&hl=en&oi=ao">researchers</a> who <a href="https://www.diversityinsport.com/lab-members">study</a> diversity, equity and inclusion in sport, we were interested in understanding what prompted such bans. Though not a surprise, we showed for the first time through an <a href="https://static1.squarespace.com/static/53e51960e4b0f38ca4081a61/t/656fa1c769a9f848e8b4c25e/1701814727449/Politics%2C+Bias%2C+and+Transgender+Athlete+Ban+%28Blind%29+%28Final%29.pdf">in-depth study</a> set to be published in the peer-reviewed <a href="https://doi.org/10.1123/jsm.2023-0137">Journal of Sport Management</a> that state-level politics and public biases against transgender people are largely to blame.</p>
<h2>Our research</h2>
<p>We collected two years of data in 2021 and 2022 on <a href="https://www.lgbtmap.org/equality-maps/youth/sports_participation_bans">states that passed</a> legislation prohibiting transgender athletes from participating in sports on teams that connect with their own gender identities. </p>
<p>To determine the political leanings of a state’s population, we <a href="https://www.ncsl.org/about-state-legislatures/state-partisan-composition">collected data</a> about the share of Republican state senators and the party affiliation of the governor.</p>
<p>Finally, we collected information about the biases people had toward transgender individuals. The data came from responses to the <a href="https://implicit.harvard.edu/implicit/takeatest.html">Project Implicit website</a>. People visiting the site can take tests aimed at measuring their biases toward different groups, including transgender people. Administrators then remove identifying information and <a href="https://osf.io/y9hiq/">make the data</a> freely available. For our study, we aggregated the responses to have transgender bias scores for each state.</p>
<h2>The politics of transgender bans</h2>
<p>States whose residents have conservative political leanings tend to have more restrictive views on civil rights issues such as <a href="https://doi.org/DOI:10.1177/1532440014524212">immigration</a>, <a href="https://doi.org/10.2105/AJPH.2020.306126">health care</a> and the use of the <a href="https://doi.org/10.1353/sof.2004.0115">death penalty</a>.</p>
<p>These patterns hold for <a href="https://doi.org/10.1177/0160323X21990839">transgender rights</a>, too.</p>
<p>In our work, we found that states with conservative-leaning legislatures such as in <a href="https://www.npr.org/2023/03/21/1164917836/wyoming-governor-calls-trans-athlete-ban-bill-draconian-and-then-allows-it-to-pa">Wyoming</a> and <a href="https://www.pbs.org/newshour/politics/supreme-court-says-transgender-girl-can-run-track-in-west-virginia-as-lawsuit-proceeds">West Virginia</a> were most likely to enact transgender athlete bans. As were states with Republican governors, such as <a href="https://www.politico.com/news/2021/06/01/desantis-transgender-sports-bill-491495">Ron DeSantis</a> in Florida and <a href="https://www.cbsnews.com/news/texas-transgender-college-athletes-bill-greg-abbott-sb-15/">Greg Abbott</a> in Texas.</p>
<p>These statewide patterns are consistent with national political actions. </p>
<p>In 2023, the Biden administration <a href="https://www.cnn.com/2023/05/12/politics/republican-governors-letter-transgender-sports-ban-title-ix/index.html">proposed a change</a> to Title IX, the federal law that bans sex discrimination at K-12 schools and colleges that receive federal funds. Under Biden’s <a href="https://www.nytimes.com/2023/04/06/us/transgender-athletes-title-ix-biden-adminstration.html">proposed changes</a>, Title IX would also ban discrimination based on sexual orientation and gender identity.</p>
<p>In response, nearly all – <a href="https://www.cnn.com/2023/05/12/politics/republican-governors-letter-transgender-sports-ban-title-ix/index.html">25 of the 26 Republican governors</a> – called on Biden to delay or withdraw the rule change. To date, Biden <a href="https://www.edweek.org/policy-politics/a-flood-of-public-feedback-has-delayed-a-title-ix-change-covering-trans-athletes-again/2023/09">has not made</a> a final decision and has delayed the change.</p>
<h2>Bias against transgender people</h2>
<p>But politics tells only part of the story.</p>
<p>We found that conservative political leanings spurred collective biases against transgender people, which in turn prompted the bans.</p>
<p>Political scientists have previously shown that politicians craft narratives and frame their arguments in ways that help shape people’s attitudes about <a href="https://doi.org/10.1111/psj.12475">social issues</a>. In fact, people will sometimes <a href="https://doi.org/10.1111/ajps.12243">adjust their perspectives</a> to align with those held by their political representatives. </p>
<p>That’s what we found.</p>
<h2>Impact on sports and athletes</h2>
<p>Biases that are prevalent in a community or state represent <a href="https://doi.org/10.1016/j.tics.2021.08.001">systemic forms</a> of oppression. Coupled with laws that limit rights, collective biases serve to <a href="https://doi.org/10.1016/j.socscimed.2015.11.010">stigmatize transgender people</a>, hurting their overall <a href="https://doi.org/10.1186/s12889-023-15856-9">health and well-being</a>. </p>
<p>The impact is far-reaching. </p>
<p>Transgender athletes face the real possibility of participating in a sport one day, only to be prohibited from doing so the next. Ending a career in sports, regardless at what age, <a href="http://csri-jiia.org/wp-content/uploads/2019/09/RA_2019_19.pdf.pdf">can harm</a> the mental health of some athletes, something only likely to be magnified given the reason for the end. </p>
<figure class="align-center ">
<img alt="Four sprinters run at an indoor track meet." src="https://images.theconversation.com/files/401062/original/file-20210517-21-km9t86.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/401062/original/file-20210517-21-km9t86.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/401062/original/file-20210517-21-km9t86.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/401062/original/file-20210517-21-km9t86.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/401062/original/file-20210517-21-km9t86.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/401062/original/file-20210517-21-km9t86.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/401062/original/file-20210517-21-km9t86.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">Bloomfield High School transgender athlete Terry Miller, second from left, wins the final of the 55-meter dash over transgender athlete Andraya Yearwood, far left, at a Connecticut girls Class S indoor track meet in 2019.</span>
<span class="attribution"><a class="source" href="https://newsroom.ap.org/detail/TransgenderAthletesHighSchool/bf19d959b3c24a53b4d315f9a26f8ddc/photo?Query=title%20AND%20ix&mediaType=photo&sortBy=arrivaldatetime:desc&dateRange=Anytime&totalCount=181&currentItemNo=2">AP Photo/Pat Eaton-Robb</a></span>
</figcaption>
</figure>
<p>Coaches and sport administrators living in conservative states might find themselves having to navigate laws affecting who can play on their teams. They can do so by partnering with <a href="https://doi.org/10.1016/j.amepre.2019.04.025">campus counselors</a> and ensuring their athletic departments are <a href="https://doi.org/10.1123/jsm.2014-0135">inclusive spaces</a>. </p>
<h2>What’s next?</h2>
<p>The links among conservative politics, collective biases against transgender people and transgender rights are unlikely to diminish any time soon. National political reporters <a href="https://www.nytimes.com/2023/04/16/us/politics/transgender-conservative-campaign.html">Adam Nagourney and Jeremy Peter</a> explained that social conservatives have targeted transgender rights as a way of galvanizing their constituents. The GOP efforts came about after planning by national conservative organizations to “harness the emotion around gender politics.”</p>
<p>Proponents of transgender inclusion <a href="https://doi.org/10.1108/978-1-80262-985-920221002">have offered counterarguments</a>, showing that transgender athletes are not a threat to women’s sports, nor have they ever been. </p>
<p>This data is important but will go only so far when combating biases. </p>
<p>Education and the chance to be around transgender people in everyday life also help <a href="https://doi.org/10.1007/s11199-011-0110-6">curb prejudice</a>. These collective factors, when combined with <a href="https://hbr.org/2011/03/using-stories-as-a-tool-of-per">compelling stories</a> about transgender inclusion in sports, may be what’s needed to overcome the biases in place.</p><img src="https://counter.theconversation.com/content/217891/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 organization that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.</span></em></p>Sports researchers learned that conservative political leanings among state legislators lead to biases against transgender athletes among voters.George B. Cunningham, UAA Endowed Professor of Sport Management, University of FloridaKelsey Garrison, PhD student, Department of Sports Management, University of FloridaLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2141362023-10-31T21:38:18Z2023-10-31T21:38:18ZStuck in the waiting room: Why women and minority groups are still underrepresented in top management<figure><img src="https://images.theconversation.com/files/549672/original/file-20230913-33750-imqte4.jpg?ixlib=rb-1.1.0&rect=40%2C0%2C4470%2C2964&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">The percentage of women at the helm of companies in North America still hovers around five per cent.
</span> <span class="attribution"><span class="source">(Shutterstock)</span></span></figcaption></figure><p>Over the past few years, there has been a great deal of research looking at how well women are represented in top companies. The findings continue to be distressing. </p>
<p>Whether in Canada or the United States, the proportion of women in top management in large organizations still hovers around five per cent. </p>
<p>Can we expect this percentage to increase over the next few years? Will today’s pool of up-and-coming female talent ensure a substantial increase in the number of female CEOs, or will other strategies be required to change the game?</p>
<p>As dean of the John Molson School of Business and a decades-long expert on the place of women in the upper echelons of the business world, we are interested in explaining the current standstill.</p>
<h2>Diversity in the C-Suite</h2>
<p>A <a href="https://www.semanticscholar.org/paper/Diversity-in-the-C-Suite%3A-The-Dismal-State-of-Among-Larcker-Tayan/192970d4859158281b752be4b76bdf7e8dc0a2c6">recent study</a> published by Stanford University professors David F. Larcker and Brian Tayan provides us with some interesting information on this subject. The aim of the study was to identify the potential for women and members of cultural communities to be appointed to CEO positions in the top 100 U.S. companies. The authors evaluated those who hold positions that report directly to a CEO. </p>
<p>The conclusions of this analysis are worrying:</p>
<ul>
<li><p>only 25 per cent of women hold such positions;</p></li>
<li><p>few women can be found in the functions that have the greatest potential for promotion, i.e. operations (15 per cent), financial services (14 per cent) and legal services (35 per cent);</p></li>
<li><p>the functions that offer fewer opportunities for promotion to CEO, according to the criteria used to select potential CEOs, are occupied to a greater extent by women (head of human resources, risk management, communications, etc.).</p></li>
</ul>
<p>The greater presence of women in these support functions illustrates <a href="https://hbr.org/2007/09/women-and-the-labyrinth-of-leadership">the leadership labyrinth</a>, i.e. the complex, dead-end detours that women face in their careers due to stereotypes, biases and family responsibilities that they continue to shoulder alone, despite better sharing of these functions with their male partners.</p>
<p>Why, after so many decades of efforts to increase female representation in decision-making bodies, do so few women manage to hold these positions? We are proposing three sources of indirect discrimination as an explanation for this.</p>
<h2>Lack of experience, a discriminatory criterion</h2>
<p><a href="https://www.spencerstuart.com/research-and-insight/predicting-ceo-success-when-potential-outperforms-experience">In a recent article</a> published by consulting firm Spencer Stuart, it was noted that the demand for experienced CEOs had almost quadrupled since the turn of the century, rising from four per cent in 1997 to 16 per cent in 2019. According to executives consulted by the firm, those in charge of selection processes assume that prior CEO experience is a predictor of the impact a candidate will have on shareholder value.</p>
<p><a href="https://www.spencerstuart.com/-/media/2019/hbr-ceo-lifecycle/hbr_ceo_lifecycle_spencerstuart.pdf">The findings of another study</a> carried out by the same firm on the life cycle of CEOs and their performance cast doubt on the assumption that there is a link between prior experience and shareholder value. </p>
<p>After analyzing the performance of 855 S&P CEOs over a 20-year period, the firm found that first-time CEOs produced a higher rate of shareholder return (TSR) than did experienced CEOs. These non-experienced CEOs had also demonstrated the advantage of staying in the job longer and having a less volatile performance overall. </p>
<p><a href="https://www.researchgate.net/publication/272806955_La_remuneration_des_dirigeants_mythes_et_recommandations">According to the results of another study</a> carried out a few years ago by Professors Michel Magnan of Concordia University and Sylvie St-Onge of HEC Montréal, less than 10 per cent of the differences in the stock market performance of the major Canadian banks can be explained by factors specific to each bank. These include the decisions and initiatives of the incumbent CEO, as well as the bank’s employees, customer base, business location and business mix.</p>
<p>The criterion of prior CEO experience, and the importance attached to it, is a factor of indirect discrimination that prevents women, members of cultural communities and young talent from having access to these positions. In addition to being discriminatory, this criterion perpetuates the status quo and limits access to such positions to a restricted group of individuals.</p>
<h2>Hiring people who look like us</h2>
<p>The concept of “cultural fit” aims to select talented individuals who are in line with the company’s culture, i.e. its values, vision, role, objectives and other elements that make up its character. </p>
<p>While using this criterion to recruit has the advantage of attracting talent who will integrate and perform quickly, it has the disadvantage of favouring the status quo and majority rule. It also means we surround ourselves with people who resemble us, whether in terms of gender, age, cultural origins or other differences <a href="https://insight.kellogg.northwestern.edu/article/cultural-fit-discrimination">that might be seen as disrupting the status quo</a>. </p>
<p><a href="https://www.talentinnovation.org/_private/assets/IDMG-ExecSummFINAL-CTI.pdf">A study</a> from the Center for Talent Innovation clearly shows that innovation thrives in an environment where leaders accept difference, are open to change and disruption, and encourage free expression.</p>
<h2>The underestimated financial value of diversity</h2>
<p>A group of researchers from Bryant University and Concordia University <a href="https://journals.aom.org/doi/abs/10.5465/AMPROC.2023.134bp">carried out an empirical study of the financial performance</a> of CEOs at the head of publicly traded U.S. companies. More than 11,600 observations were made each year over a 15-year period (1998-2013). </p>
<p>They found that women of colour and white women outperformed men of colour, who outperformed white men. According to the authors, these results can be explained by the fact that from a very young age, people from minority backgrounds are told by those around them that they need to develop resilience, and that if they want to succeed, they need to be smarter and do better than anyone else. </p>
<h2>More human leadership</h2>
<p>In a world where volatility, uncertainty, complexity and ambiguity prevail, the leadership qualities that are appropriate to such a context should guide selection processes. These qualities — agility, adaptability, empathy, humility — can be found in both men and women. These are what we call “soft skills.” </p>
<p><a href="https://hbr.org/2022/07/the-c-suite-skills-that-matter-most">According to a study published in August 2022</a>, the quest for these qualities has become increasingly important in job descriptions for senior management positions over the past decade. Prioritizing the qualities that allow us to identify the best candidates is the only way we will ensure a level playing field for women and men alike. </p>
<p>Companies can benefit from recognizing the importance of diversity in talent and leadership styles. By promoting the best people to positions of power, companies will become more efficient and more humane.</p><img src="https://counter.theconversation.com/content/214136/count.gif" alt="La Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Les auteurs ne travaillent pas, ne conseillent pas, ne possèdent pas de parts, ne reçoivent pas de fonds d'une organisation qui pourrait tirer profit de cet article, et n'ont déclaré aucune autre affiliation que leur organisme de recherche.</span></em></p>After decades of efforts to increase female representation in corporate decision-making bodies, few women are managing to take the reins of power.Louise Champoux-Paillé, Cadre en exercice, John Molson School of Business, Concordia UniversityAnne-Marie Croteau, Dean, John Molson School of Business, Concordia UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2148122023-10-31T12:33:47Z2023-10-31T12:33:47ZWorkplace discrimination saps everyone’s motivation − even if it works in your favor<figure><img src="https://images.theconversation.com/files/554087/original/file-20231016-17-idklp6.jpg?ixlib=rb-1.1.0&rect=0%2C36%2C4851%2C3554&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">If your boss is biased, this is a logical response.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/businessman-sleeping-on-computer-keyboard-royalty-free-image/85406503">Robert Daly/OJO Images via Getty Images</a></span></figcaption></figure><p>When people work for discriminatory managers, they put in less effort. That’s true both when managers are biased against them and when they’re biased in their favor, according to <a href="https://rdcu.be/dmIdQ">a new paper</a> that Nicholas Heiserman of Oklahoma State University <a href="https://sc.edu/study/colleges_schools/artsandsciences/sociology/our_people/faculty_staff_directory/simpson_brent.php">and I</a> have published in the journal Nature Human Behaviour. </p>
<p>To demonstrate this, we placed nearly 1,200 research participants in several experiments designed to mimic work settings, where they and other “workers” made decisions about how much effort to dedicate to a task. </p>
<p>In some experiments, we had participants complete number searches – by counting how many times “3” appeared in a large table of numbers, for example. The more searches a participant completed, the higher their effort was rated. Participants, working in pairs or in small groups, were told that their manager would award a bonus to one person based on how many number searches the workers completed. </p>
<p>To create a discriminatory situation, participants were told that there were two types of employees: blue and red. Participants were always assigned to be blue. One-third of the participants were told that the manager had a bias against blue employees, while another third were told that the manager was biased in their favor. The rest didn’t receive any information one way or the other.</p>
<p>We found that those workers who knew their managers discriminated – whether it was for them or against them – completed fewer number searches than participants in the control group. </p>
<p>By measuring workers’ expectations that they would receive a bonus, our experiments also help show that discrimination reduces work productivity by separating effort from rewards. </p>
<p>This makes intuitive sense: If you know your boss is biased against people like you, you’ll have less incentive to work hard, since you know you’re unlikely to get promoted regardless. Similarly, if your boss is biased in favor of people like you, you’ll probably get promoted anyway. So, again, why work hard?</p>
<h2>Why it matters</h2>
<p>It’s well established that workplace discrimination leads to <a href="https://doi.org/10.1086/511799">reduced earnings</a> and <a href="https://doi.org/10.3389/fpsyg.2015.01400">advancement opportunities</a> for members of disadvantaged groups. </p>
<p>But our results suggest that it can lower productivity of all workers, even those advantaged by it – which means discrimination may hurt firms’ bottom lines more than has been assumed. </p>
<p>Another of our key findings helps explain why the effects of discrimination on work effort can worsen over time. Specifically, we found that even though working for a discriminatory boss made everyone put in less effort, the disadvantaged showed the largest decline. </p>
<p>We suspect this could lead to a vicious cycle, where targets of discrimination respond by putting in less effort than advantaged workers. In turn, their managers may come to see them as lazier, less competent or less deserving of promotions – which can strengthen their original biases.</p>
<p>To test this, we ran an additional study with participants who had managerial experience. We showed them the work effort of two groups of participants from our experiments: one group that had been discriminated against, and one that benefited from discrimination against others. The latter group had higher productivity. </p>
<p>We labeled these groups generically as “red types” and “blue types,” and while the managers knew that one group had put in more effort, they didn’t know discrimination was the reason why. </p>
<p>We found that managers readily stereotyped both groups, perceiving members of the advantaged group as warmer and much more competent. Further, they said they would strongly prefer to hire, work with, promote and give bonuses to members of the advantaged category. </p>
<p>These findings show how discrimination can lead to behavior by employees that strengthens the negative stereotypes underlying the original act of discrimination, or even spread discriminatory stereotypes to new managers. </p>
<h2>What’s next</h2>
<p>Studying discrimination based on invented categories in simulated work environments can help us understand the basics of how it works, but it ignores differences in how bias operates when it comes to, for instance, race versus gender, or sexuality versus parental status. An important goal for future research is to better understand how the processes we observe play out for these real-world bases of discrimination. </p>
<p>For instance, following a <a href="https://doi.org/10.1093/qje/qjx006">related study</a>, future research might measure racial biases of managers in organizations and the productivity of employees who work for them. Based on our research, we would expect employees whose managers are racially biased to be less productive than employees whose managers aren’t.</p>
<p>But we may expect different effects if, rather than racial discrimination, we studied the <a href="https://doi.org/10.1111/soc4.12031">well-established</a> pattern of discrimination against mothers in the workplace. That’s because, as we have shown in <a href="https://doi.org/10.1016/j.ssresearch.2021.102642">our prior work</a>, some mothers don’t interpret clearly biased treatment of them in the workplace as discriminatory. So what happens when people work for biased managers but don’t recognize it? That’s an important question to address in future work. </p>
<p><em>The <a href="https://theconversation.com/us/topics/research-brief-83231">Research Brief</a> is a short take on interesting academic work.</em></p><img src="https://counter.theconversation.com/content/214812/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Brent Simpson receives funding from the National Science Foundation and Army Research Office. </span></em></p>Having a biased manager lowers productivity across the board – even for workers who aren’t targeted.Brent Simpson, Professor of Sociology, University of South CarolinaLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2149432023-10-08T21:30:02Z2023-10-08T21:30:02ZRomantic heroes or ‘one of us’ – how we judge political leaders is rarely objective or rational<p>Given the presidential style of modern politics, the intense media focus on party leaders is unavoidable. But this involves a degree of artifice. New Zealanders don’t vote directly for a prime minister, they vote for their preferred <a href="https://vote.nz/2023-general-election/about/2023-general-election/?">party and electorate candidate</a>. </p>
<p>Technicalities aside, though, party leaders <a href="https://www.routledge.com/The-Selection-of-Political-Party-Leaders-in-Contemporary-Parliamentary/Pilet-Cross/p/book/9781138187573">play a key role</a> in shaping their party’s policies and soliciting public support. The upside of the attention they receive, therefore, is that voters get to scrutinise before they “buy”. </p>
<p>That’s especially important for undecided or swing voters. Not only can they compare policies, they can also examine each leader’s strengths and weaknesses, and gauge what values guide their approach.</p>
<p>Being head of state is a hugely challenging role – not least because leaders fundamentally get results <a href="https://www.amazon.com.au/Oxford-Management-Readers-Keith-Grint/dp/0198781814">through mobilising collective effort</a>. If no one’s following, there is no leadership.</p>
<p>This is <a href="https://www.amazon.com.au/Leadership-Theory-Practice-Peter-Northouse/dp/1544397569/ref=sr_1_3?">different from management</a>, which largely revolves around detailed planning and then implementing and monitoring progress toward goals. </p>
<p><a href="https://www.hup.harvard.edu/catalog.php?isbn=9780674518582">Leadership</a>, however, involves connecting with people’s values and needs, and helping them make sense of events. It entails crafting a vision for the future and formulating credible strategies to achieve that. </p>
<p>It involves the capacity to make wise decisions, and role modelling what it means to be a person of good character. While managerial competency still matters, being a prime minister demands far more. </p>
<p>That said, trying to objectively evaluate a potential leader is not easy.</p>
<h2>Favouring our own team</h2>
<p>Humans strongly favour those they view as being “one of us”. A large body of <a href="https://link.springer.com/chapter/10.1007/978-3-319-29869-6_1">research shows</a> people trust, respect, support, care for and are more influenced by those they feel an affinity with. This sense of a shared identity might be based on common demographic characteristics, or shared interests and values.</p>
<p>Dedicated sports fans illustrate this well. Individually and collectively, they back their team no matter what. They wear team colours, idealise team members, mock opponents and boo referees – even if the ref is right. </p>
<p>The same socio-psychological forces are at play in the political domain. </p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/controlling-the-political-narrative-is-key-to-winning-the-nz-election-no-easy-task-for-chris-hipkins-213533">Controlling the political narrative is key to winning the NZ election – no easy task for Chris Hipkins</a>
</strong>
</em>
</p>
<hr>
<p>The party faithful are unlikely to offer an objective view. They will likely overestimate the strengths and underestimate the weaknesses of their party’s policies and its leaders (and do the opposite when evaluating opposing parties).</p>
<p>This is clearly not helpful for undecided or swing voters. But even beyond partisan influences, determining what constitutes good leadership is a more vexed issue than we might imagine. </p>
<p>While people often hold strong views, the actual evidence about what constitutes “good” leadership is quite diverse and complex. </p>
<h2>Fantasies and realities</h2>
<p>There are many different theories, but researchers generally agree that “<a href="https://www.amazon.com/Ethics-Heart-Leadership-Joanne-Ciulla/dp/1440830657/ref=sr_1_1?">good leadership</a>” is both ethical and effective. But people can often ignore those considerations when evaluating someone in (or aspiring to) a leadership role. </p>
<p>Subconsciously, we are inclined to judge leaders according to our own personal theories of leadership. This “<a href="https://www.sciencedirect.com/science/article/pii/S1048984317304988">implicit</a>” bias is typically shaped by the kinds of behaviours <a href="https://journals.sagepub.com/doi/abs/10.1177/1548051820931243">role-modelled by the authority figures</a> we were exposed to early in life. </p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/liz-truss-is-now-a-case-study-in-poor-leadership-192554">Liz Truss is now a case study in poor leadership</a>
</strong>
</em>
</p>
<hr>
<p>A very strict parenting style, for example, which a child finds reassuring rather than restrictive, can lead them in later life to favour command-oriented or even authoritarian leaders.</p>
<p>Research indicates that even in democracies, about one-third of the population <a href="https://www.cambridge.org/core/books/authoritarian-dynamic/7620B99124ED2DBFC6394444838F455A">favours</a> that kind of traditional “strong man” leadership style. </p>
<p>But <a href="https://www.elgaronline.com/monobook/9781784716783.xml">scholars have long argued</a> such leaders tend to be intolerant, oppressive, punitive, lacking in empathy and prone to bullying. They may resist being held to account for their actions, arrogantly believing they know best.</p>
<h2>Romantic attachments</h2>
<p>Popular culture and media narratives are other <a href="https://www.e-elgar.com/shop/gbp/leadership-popular-culture-and-social-change-9781785368967.html">important influences</a> on people’s ideas about leadership. In books, TV shows and movies, leaders are often depicted as heroic, larger-than-life characters with the capacity to save others, even the world. In the business media, it’s often implied CEOs have somehow turned a company around single-handedly. </p>
<p>Researchers call this the “<a href="https://www.jstor.org/stable/2392813">romance</a>” of leadership – a tendency to overstate what leaders can actually do, and to blame them when they fail to meet unrealistic expectations.</p>
<p>Indeed, no matter how skilled and dedicated, leaders are inevitably flawed, just like the rest of humanity. Nor are they omnipotent. New Zealand is a small and remote trading nation in an interconnected world, not a superpower or totalitarian state. There are many things its prime minister cannot control.</p>
<p>In that sense, an ability to <a href="https://www.hup.harvard.edu/catalog.php?isbn=9780674518582">manage expectations</a> is an indicator of good leadership. Having the personal integrity to avoid making unrealistic promises is what serves democracy. Offering <a href="https://www.amazon.com.au/Allure-Toxic-Leaders-Destructive-Politicians_and/dp/0195312007">false hope</a> is not good leadership, it’s more like what con artists and charlatans hungry for power do. </p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/some-leaders-only-want-to-hear-the-good-news-politicians-tell-us-how-political-careers-can-end-182590">'Some leaders only want to hear the good news': politicians tell us how political careers can end</a>
</strong>
</em>
</p>
<hr>
<h2>Check your bias</h2>
<p>Overall, the evidence is that we’re not very rational or objective when it comes to evaluating leaders. Even if we’re not swayed by ideological factors, our personal experience and a romantic view of leadership can unconsciously cloud our judgment.</p>
<p>But there are some things we can do to help us make a more informed and balanced judgement. Firstly, we can try to step back and reflect on our own biases and <a href="https://journalofleadershiped.org/jole_articles/challenging-your-implicit-leadership-theory/">assumptions about leadership</a>. (You can even <a href="https://implicit.harvard.edu/implicit/takeatest.html">test your own bias</a> on a number of issues with the Project Implicit online resource.)</p>
<p>Secondly, look for indicators of behaviours associated with good leadership. Many of these are the same <a href="https://www.amazon.com/Developing-Leadership-Character-Mary-Crossan/dp/1138825670/ref=sr_1_1?">character virtues</a> we’d admire in anyone: integrity, fair-mindedness, the determination to do their best, confidence (but not arrogance), being accountable for their actions, and empathy and respect for others. These are vital foundations for good leadership.</p>
<p>And thirdly, look for actual evidence of leadership skills. Being prime minister is complex and challenging. It demands an ability to address serious issues in a serious-minded way. </p>
<p>Good leaders are not glib, superficial or unable to answer valid and reasonable questions. Consequently, a good leader <a href="https://www.hup.harvard.edu/catalog.php?isbn=9780674518582">may not tell you what you want to hear</a>. But if they encourage us all to <a href="https://www.sciencedirect.com/science/article/pii/S0263237310000277">address difficult realities</a>, that’s to be admired, not condemned.</p>
<p>As to whether that’s enough to win your vote, only you can be the judge of that.</p><img src="https://counter.theconversation.com/content/214943/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Suze Wilson 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>Personal bias, upbringing and even popular dramas can influence the way we evaluate political leadership. As election day nears, how might we make more balanced judgments?Suze Wilson, Senior Lecturer, School of Management, Massey UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2141982023-09-27T21:34:26Z2023-09-27T21:34:26ZHidden in plain sight: Women face subtle forms of discrimination and bias in the workplace<figure><img src="https://images.theconversation.com/files/550442/original/file-20230926-19-ovbqmb.jpg?ixlib=rb-1.1.0&rect=0%2C8%2C5472%2C3628&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">The subtler, more insidious forms of discrimination that women face at work often go unnoticed.</span> <span class="attribution"><span class="source">(Shutterstock)</span></span></figcaption></figure><iframe style="width: 100%; height: 100px; border: none; position: relative; z-index: 1;" allowtransparency="" allow="clipboard-read; clipboard-write" src="https://narrations.ad-auris.com/widget/the-conversation-canada/hidden-in-plain-sight-women-face-subtle-forms-of-discrimination-and-bias-in-the-workplace" width="100%" height="400"></iframe>
<p>Gender discrimination remains a <a href="https://www.pewresearch.org/short-reads/2017/12/14/gender-discrimination-comes-in-many-forms-for-todays-working-women/">pervasive issue</a> in <a href="https://www.canada.ca/en/employment-social-development/corporate/reports/women-symposium.html">the workplace</a>. While obvious cases of discrimination against women — like <a href="https://www.thestar.com/podcasts/this-matters/a-toronto-police-officer-shares-her-story-of-surviving-workplace-sexual-harassment/article_1a600227-7864-5388-bdcf-521010066b89.html">sexist comments</a> or <a href="https://www.cbc.ca/player/play/2199225923548">the systematic underpayment of women</a> — dominate headlines, there are subtler, more insidious forms of discrimination that often go unnoticed.</p>
<p>Take Kelly, for example, a seasoned marketing manager we recently interviewed as part of a workplace discrimination project. Kelly had diligently worked towards a promotion, only to witness her junior colleague, Mark, receive it instead. This led her to wonder if Mark genuinely outperformed her, or if there was something more nefarious at play.</p>
<p>Kelly’s quandary isn’t unique. It reflects a pervasive, subtle challenge faced by women in many fields: incidents tinged with potential gender bias, yet ambiguous enough to defy clear categorization as discrimination.</p>
<p>It’s easy to condemn blatant discrimination because of how obvious it is. But discrimination doesn’t always reveal itself so openly; instead, <a href="https://www.nytimes.com/2017/07/19/us/was-that-racist.html">it can be a spectre</a> looming uncertainly in the background. </p>
<h2>Examining ambiguous incidents</h2>
<p><a href="https://doi.org/10.15195/v10.a18">Our recent research</a> aimed to investigate women’s experiences of ambiguous incidents in the workplace. Seeking to understand the issue from multiple angles, we conducted interviews, a survey and an experiment.</p>
<p>The project uncovered myriad tales of women grappling with incidents that might have been driven by bias, but were cloaked in uncertainty. Their stories encompassed a wide spectrum of experiences, ranging from daily microaggressions, such as being ignored during meetings, to significant career milestones, like missing out on promotions. </p>
<figure class="align-center ">
<img alt="A forlorn-looking woman stands with her arms folded while two men shake hands in the background" src="https://images.theconversation.com/files/550439/original/file-20230926-15-1v6b7y.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/550439/original/file-20230926-15-1v6b7y.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=460&fit=crop&dpr=1 600w, https://images.theconversation.com/files/550439/original/file-20230926-15-1v6b7y.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=460&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/550439/original/file-20230926-15-1v6b7y.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=460&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/550439/original/file-20230926-15-1v6b7y.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=578&fit=crop&dpr=1 754w, https://images.theconversation.com/files/550439/original/file-20230926-15-1v6b7y.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=578&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/550439/original/file-20230926-15-1v6b7y.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=578&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">The experiences of women in the workplace range from daily microaggressions, such as being ignored during meetings, to significant career milestones, like missing out on promotions.</span>
<span class="attribution"><span class="source">(Shutterstock)</span></span>
</figcaption>
</figure>
<p>Most of the women we interviewed wrestled more with ambiguous incidents than with overt discrimination. As Kelly put it: </p>
<blockquote>
<p>“I think I would feel better if it was overtly gender discrimination, because at least you would feel somewhat validated in your perception, whereas you always question, like, maybe I’m not seeing things right, maybe I’m biased.”</p>
</blockquote>
<p>Like Kelly, 74 per cent of the women we surveyed reported that they had struggled with such ambiguities in the past year. Only 64 per cent said they had faced clear-cut discrimination. These aren’t just numbers; they represent the silent battles and moments of self-doubt that many women experience.</p>
<h2>Responding to possible discrimination</h2>
<p>Following ambiguous incidents, many women reported feeling confused or frustrated, often ruminating over their experiences and struggling to make sense of them. But, as we found, ambiguous incidents had more than just emotional impacts. </p>
<p>We designed an experiment in which participants were exposed to the same discrimination incident, but at different levels of ambiguity. Some participants experienced the incident as clear-cut discrimination, whereas others experienced it as ambiguous. </p>
<p>The experiment revealed that when a situation is clearly discriminatory, women are more likely to turn outwards by speaking to human resources, consulting with supervisors or seeking advice from diversity and inclusion groups. This sort of action not only addresses the issue at hand, but also sets the stage for organizational change.</p>
<p>But when an incident is ambiguous, women tend to turn inwards. They try to adopt a more formal communication style, work harder or draw more attention to their achievements. While this may help them navigate discrimination in the short term, it does little to catalyze the kind of systemic change necessary to foster gender equality.</p>
<h2>A call to leaders and allies</h2>
<p>What can leaders and allies do to help? </p>
<p>First, we all need to shatter the silence that surrounds these incidents. Ambiguity thrives when communication is stifled. Creating an environment where whispers of concern are welcomed, not shunned, is paramount. This goes beyond just having an <a href="https://doi.org/10.5465/amj.2007.26279183">open-door policy</a>; it’s about building trust so that people know those doors lead to empathetic listeners.</p>
<p><a href="https://hbr.org/2022/11/7-ways-to-practice-active-allyship">Allies can also play a powerful role</a>. When someone stands up and acknowledges these subtle biases, it doesn’t just validate feelings, but also builds bridges. When colleagues and managers notice ambiguous discrimination, they should take the initiative to engage in private discussions with the affected women. A simple acknowledgement or private conversation can shift the narrative from doubt to trust. </p>
<figure class="align-center ">
<img alt="Two women have a conversation at a desk. One woman has her back to the camera." src="https://images.theconversation.com/files/550440/original/file-20230926-15-fsn084.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/550440/original/file-20230926-15-fsn084.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/550440/original/file-20230926-15-fsn084.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/550440/original/file-20230926-15-fsn084.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/550440/original/file-20230926-15-fsn084.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/550440/original/file-20230926-15-fsn084.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/550440/original/file-20230926-15-fsn084.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">When colleagues and managers notice ambiguous discrimination, they should take the initiative to engage in private discussions with affected women.</span>
<span class="attribution"><span class="source">(Shutterstock)</span></span>
</figcaption>
</figure>
<p>However, it’s essential to exercise caution. While it’s important to listen, it’s equally critical to distinguish between unintentional missteps and genuine bias. Colleagues and managers must take concerns seriously without unfairly penalizing people whose actions were ambiguous, but not biased. </p>
<p>To navigate this fine line, we must adopt a prudent approach. This involves seeking multiple perspectives, conducting thorough investigations and thoughtfully considering the context in which incidents occurred. </p>
<p>Lastly, as women start sharing their ambiguous experiences, their managers and colleagues should look for recurring themes. A single event may be an aberration, but a pattern is a cause for alarm. It signals systemic problems that require attention. </p>
<p>It’s vital to recognize that, in the journey towards diversity and inclusion, it’s not just the visible mountains we need to climb. Often, it’s the foggy valleys of ambiguity that prove the most challenging to traverse.</p><img src="https://counter.theconversation.com/content/214198/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Laura Doering receives funding from the Social Sciences and Humanities Research Council of Canada.</span></em></p><p class="fine-print"><em><span>András Tilcsik receives funding from the Social Sciences and Humanities Research Council of Canada.
</span></em></p><p class="fine-print"><em><span>Jan Doering receives funding from the Social Sciences and Humanities Research Council of Canada.</span></em></p>While blatant discrimination is easy to condemn because of how obvious it is, there are subtler, more insidious forms that also need to be rooted out.Laura Doering, Associate Professor of Strategic Management, University of TorontoAndrás Tilcsik, Professor of Strategic Management, University of TorontoJan Doering, Assistant Professor of Sociology, University of TorontoLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2129462023-09-27T12:24:22Z2023-09-27T12:24:22ZHarassment and abuse perceived to harm poor women less − new research finds a ‘thicker skin’ bias<p>People think sexual harassment and domestic abuse are less harmful for women in poverty than for higher-income women, according to <a href="https://doi.org/10.1016/j.jesp.2023.104472">four studies</a> involving 3,052 Americans conducted by <a href="https://scholar.google.com/citations?user=McvKSycAAAAJ&hl=en&oi=ao">my</a> <a href="https://scholar.google.com/citations?user=hNlMsXkAAAAJ&hl=en&oi=ao">colleagues</a> and <a href="https://scholar.google.com/citations?hl=en&user=H-H6_qYAAAAJ">me</a>. We also found that people believe women in poverty require less help and support when experiencing these kinds of sexual misconduct. </p>
<p>My research partners and I recruited participants of different ages, genders and incomes. We asked them to read about either a low-income woman or a high-income woman who was dealing with workplace sexual harassment or intimate partner abuse. Then we had participants rate how distressing these instances would be for the woman. </p>
<p>The harassment events described inappropriate behavior from a co-worker, such as sexual comments and unwanted advances, while domestic abuse events included threats, demeaning comments and physical violence from the woman’s partner. In some of the studies, participants also rated how much social support or bystander intervention would be necessary for these events.</p>
<p>Our participants rated the harassment and abuse events as less upsetting for the lower-income woman than for the higher-income woman. They also thought the lower-income woman would need less emotional support from friends and family and less help from bystanders than the higher-income woman. On average, participants thought she needed only <a href="https://doi.org/10.1016/j.jesp.2023.104472">85% as much help</a> as her higher-income counterpart. </p>
<p>The result was the same whether the woman was white, Black, East Asian or Latina. Both low- and high-income study participants shared this pattern of judgment – as did male and female participants.</p>
<h2>Why it matters</h2>
<p>There is no data that shows lower-income women are less affected by gender-based violence – in fact, there is evidence they are often <a href="https://doi.org/10.1891/0886-6708.20.6.625">more affected</a>.</p>
<p>Women in poverty are <a href="https://nwlc.org/resource/out-of-the-shadows-an-analysis-of-sexual-harassment-charges-filed-by-working-women">more likely to experience sexual harassment</a> and <a href="https://doi.org/10.1007/s10896-018-0019-8">domestic abuse</a> – and <a href="https://doi.org/10.5070/L3262045668">have more difficulty finding support</a> after experiencing sexual misconduct. Our research suggests that stereotypes about toughness may contribute to the neglect low-income women encounter when they seek help after violence. </p>
<p>It isn’t that study participants didn’t like the low-income woman. In fact, in our studies, participants rated the low-income woman as friendlier and warmer than the higher-income woman. But liking the low-income woman didn’t prevent participants from thinking the harassment and abuse would be less harmful for her. </p>
<p>Such perceptions may have wide-ranging consequences. For example, low-income women may not receive the care they need from those around them. They also may be disproportionately neglected by those in powerful positions, such as human resources managers and police investigating domestic abuse. </p>
<p>Biased perceptions may help explain why lower-income women <a href="https://doi.org/10.5070/L3262045668">encounter more barriers in the legal system</a>. </p>
<p>Already, the neglect of low-income women has been effectively part of U.S. federal workplace law based on <a href="https://brooklynworks.brooklaw.edu/blr/vol70/iss3/1/">several rulings from courts hearing sexual harassment claims</a>. For example, in the 1995 case <a href="https://scholar.google.com/scholar_case?case=6933014541839038196&q=Gross+v.+Burggraf&hl=en&as_sdt=40000006&as_vis=1">Gross v. Burggraf</a>, the court ruled that sexually harassing behaviors in a “white collar” workplace do not necessarily qualify as harassment in “blue collar” contexts like construction sites. </p>
<p>This logic echoes our study participants’ judgments – and also partially explains why low-income women have spoken out about being <a href="https://www.nytimes.com/2017/12/20/us/the-metoo-moment-blue-collar-women-ask-what-about-us.html">sidelined by the #MeToo movement</a>.</p>
<h2>What other research is being done</h2>
<p>Our research fits with a <a href="https://doi.org/10.1002/ejsp.2843">growing</a> <a href="https://doi.org/10.1016/j.jesp.2017.09.002">body</a> <a href="https://doi.org/10.1037/xap0000442">of work</a> examining beliefs around experiencing adversity. People seem to widely <a href="https://doi.org/10.1177/1368430215625781">endorse the idea</a> “what doesn’t kill you makes you stronger.” Individuals who have experienced past hardship, such as women experiencing financial difficulties, are perceived by others to have grown a “<a href="https://doi.org/10.1017/bpp.2020.33">thicker skin</a>,” making them less affected by new negative events.</p>
<p>Our findings show this kind of bias exists for low-income women – and highlight the need for strategies to counteract this biased belief.</p>
<p><em>The <a href="https://theconversation.com/us/topics/research-brief-83231">Research Brief</a> is a short take on interesting academic work.</em></p><img src="https://counter.theconversation.com/content/212946/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Nathan Cheek received funding from the National Science Foundation. </span></em></p>While women in poverty are more likely to experience sexual harassment and domestic abuse than higher-income women, people assume it is less distressing for them.Nathan Cheek, Assistant Professor of Psychological Sciences, Purdue UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2084232023-09-22T12:29:56Z2023-09-22T12:29:56ZBiases against Black-sounding first names can lead to discrimination in hiring, especially when employers make decisions in a hurry − new research<figure><img src="https://images.theconversation.com/files/549130/original/file-20230919-23-y3ipbc.jpg?ixlib=rb-1.1.0&rect=0%2C116%2C5301%2C2563&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">What role will race play in determining who gets the job?</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/everyones-grabbed-their-easiest-prep-tool-royalty-free-image/1174452924?phrase=hiring+job+candidates&adppopup=true">Cecilie_Arcurs/E+ via Getty Image</a></span></figcaption></figure><p>Because names are among the first things you learn about someone, they can influence first impressions. </p>
<p>That this is particularly true for names associated with Black people came to light in 2004 with the release of a study that found employers <a href="https://doi.org/10.1257/0002828042002561">seeing identical resumes</a> were 50% more likely to call back an applicant with <a href="https://abcnews.go.com/2020/top-20-whitest-blackest-names/story?id=2470131">stereotypical white names like Emily or Greg</a> versus applicants with names like Jamal or Lakisha.</p>
<p>I’m a <a href="https://scholar.google.com/citations?user=WJe3b0UAAAAJ&hl=en&oi=sra">behavioral economist who researches discrimination in labor markets</a>. In a <a href="https://dx.doi.org/10.2139/ssrn.4490163">study based on a hiring experiment</a> I conducted with another economist, <a href="https://scholar.google.com/citations?hl=en&user=vyGCfDoAAAAJ">Rulof Burger</a>, we found that participants systematically discriminated against job candidates with names they associated with Black people, especially when put under time pressure. We also found that white people who oppose affirmative action discriminated more than other people against job candidates with distinctly Black names, whether or not they had to make rushed decisions.</p>
<h2>Detecting racial biases</h2>
<p>To conduct this study, we recruited 1,500 people from all 50 U.S. states in 2022 to participate in an online experiment on <a href="https://prolific.com">Prolific</a>, a survey platform. The group was nationally representative in terms of race and ethnicity, age and gender.</p>
<p>We first collected data on their beliefs about the race and ethnicity, education, productivity and personality traits of people with six names picked from a pool of 2,400 workers whom we hired in an early stage of our experiment for a transcription task. Data from these individual responses made it possible for us to categorize how they perceived the candidates.</p>
<p>We found that the names of workers perceived as Black, such as Shanice or Terell, were more likely to elicit negative presumptions, such as being less educated, productive, trustworthy and reliable, than people with either white-sounding names, such as Melanie or Adam, or racially ambiguous names, such as Krystal or Jackson.</p>
<p>We were specifically studying discrimination against Black people, so we did not include names in this experiment that are frequently associated with Hispanics or Asians. </p>
<p>Participants were next presented with pairs of names and were told they could earn money for selecting the worker who was more productive in the transcription task. The chance that they would choose job candidates they perceived to be white because of their names was almost twice as high than if they thought the candidates to be Black. This tendency to discriminate against people with Black-sounding names was greatest among men, people over 55, whites and conservatives.</p>
<p>Educational attainment, the level of racial diversity in the participants’ ZIP codes or whether they had personally hired anyone before didn’t influence their apparent biases. </p>
<p><iframe id="cju7c" class="tc-infographic-datawrapper" src="https://datawrapper.dwcdn.net/cju7c/3/" height="400px" width="100%" style="border: none" frameborder="0"></iframe></p>
<h2>Rushing can cause more discriminatory behavior</h2>
<p>Most real-world hiring managers spend <a href="https://careers.workopolis.com/advice/employers-view-resumes-for-fewer-than-11-seconds">less than 10 seconds</a> reviewing each resume during the initial screening stage. To keep up that swift pace, they may resort to using mental shortcuts – including racial stereotypes – to assess job applications.</p>
<p>We found that requiring the study participants to select a worker within only 2 seconds led them to be 25% more likely to discriminate against candidates with names they perceived as Black-sounding. Similar patterns of biased decision-making under time pressure have been documented in the context of <a href="https://doi.org/10.1037/0022-3514.92.6.1006">police shootings</a> and <a href="https://doi.org/10.1177/0022146512445807">medical decisions</a>.</p>
<p>However, making decisions more slowly is not a panacea. </p>
<p>We found that the most important factor for whether more deliberate decisions reduce discrimination was a participant’s view on <a href="https://www.law.cornell.edu/wex/affirmative_action">affirmative action</a> – the consideration of race in a workforce or student body to ensure that their share of people of color is roughly proportionate to the general public or a local community. </p>
<p>White participants who opposed affirmative action were more than twice as likely to select an applicant with a white-sounding name compared with applicants perceived as Black – whether or not they had to make the simulated hiring decision in a hurry.</p>
<p>By contrast, giving white participants who favor affirmative action unlimited time to choose a name from the hiring list reduced discrimination against the job candidates with names they perceived as Black-sounding by almost half. The data showed that this decline had to do with people basing their decision more on their perceptions of a worker’s performance, rather than relying on mental shortcuts based on their perceived race.</p>
<p>We assessed the participants’ views on affirmative action by doing a survey at the end of this experiment.</p>
<h2>Discrimination hasn’t gone away</h2>
<p>A study published in 2021 <a href="https://www.nber.org/papers/w29053">suggested that hiring discrimination</a> based on Black-souding names had declined, although discriminatory practices remained high in some customer-facing lines of work, such as auto sales or retail. </p>
<p>Other research has suggested that once people learn more about someone, the <a href="https://doi.org/10.1086/722093">discriminatory influence that a name might have</a> begins to fade. Yet, other studies have indicated that racial biases can make the <a href="https://doi.org/10.1257/pandp.20231114">interactions needed for this learning process less likely</a>. For example, racial biases may lead employers to refrain from interviewing – or hiring – a job candidate of color in the first place.</p>
<p>There is ample evidence that people of color face discrimination in many important domains beyond employment, including finding <a href="https://doi.org/10.1257/app.20160213">housing</a> or <a href="https://doi.org/10.1093/rfs/hhac029">obtaining loans</a>.</p>
<p>Our results suggest that slowing down the initial assessment of applicants can be a first step toward reducing this type of discrimination.</p><img src="https://counter.theconversation.com/content/208423/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Martin Abel 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>People who object to affirmative action were more likely to discriminate against job candidates with Black-sounding names than those who supported it, whether or not they had to rush.Martin Abel, Assistant Professor of Economics, Bowdoin CollegeLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2097712023-08-17T12:35:12Z2023-08-17T12:35:12ZPotentially faulty data spotted in surveys of drug use and other behaviors among LGBQ youth<figure><img src="https://images.theconversation.com/files/541296/original/file-20230804-26-63jilc.jpg?ixlib=rb-1.1.0&rect=202%2C166%2C7737%2C5130&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">A new study found that youth were providing extreme or untruthful responses to CDC surveys on LGBQ student health. </span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/female-hands-of-a-student-taking-a-test-royalty-free-image/1305362771?phrase=students+taking+a+survey&adppopup=true">FG Trade/E+ via Getty Images</a></span></figcaption></figure><p>Federal data on LGBQ student health <a href="https://doi.org/10.1111/cdev.13957">contain a significant amount of potentially exaggerated or untruthful responses</a>, raising questions about how they might skew people’s understanding of risky behavior among teens. These inaccuracies affect some responses more than others. That’s according to an analysis my colleagues and I did of high school surveys administered by the Centers for Disease Control and Prevention, better known as the CDC.</p>
<p>Without accounting for this invalid data, the CDC results suggest that for every heterosexual boy who uses steroids, three LGBQ boys use steroids. After accounting for the invalid data, neither group is shown to use steroids more. In contrast, disparities for being bullied or considering suicide were not affected by potentially invalid data.</p>
<p>Over 12,800 high school students during the 2018-2019 school year reported whether they identified as LGBQ – that is, lesbian, gay, bisexual or questioning – or heterosexual on the national <a href="https://www.cdc.gov/healthyyouth/data/yrbs/data.htm">Youth Risk Behavior Survey</a>. They also responded to items related to their health and well-being. </p>
<p>We first estimated what the risk disparities between LGBQ and heterosexual youth were before accounting for potentially invalid data. We then used a machine-learning algorithm to detect response patterns that suggested when youth were providing extreme or untruthful responses.</p>
<p>For example, we treated their responses with suspicion if they reported eating carrots four or more times every day and said they were impossibly tall. That means we gave less weight to their responses when we re-estimated all of the disparities. We then saw how the disparities changed after the potentially invalid responses were taken into account.</p>
<p>After accounting for invalid data, disparities in drug use – including steroids – injected drugs, cocaine, ecstasy and pain medication without a prescription were not as pronounced. LGBQ boys appeared to use injected drugs four times as often as heterosexual boys. But after accounting for the likely invalid data, neither group was more likely to use injected drugs. </p>
<p>Yet, while some outcomes were susceptible to invalid data, others were not. For example, LGBQ boys and girls were about twice as likely to be bullied at school and two to three times as likely to consider suicide. This shows that not all outcomes are equally affected by invalid data. </p>
<h2>Why it matters</h2>
<p>The Youth Risk Behavior Survey provides vital information on the health and behaviors of high school students. It informs research regarding <a href="https://www.cdc.gov/healthyyouth/data/yrbs/yrbs_data_summary_and_trends.htm">teen sexual behaviors, drug use and suicide risk</a>.</p>
<p><a href="https://doi.org/10.1111/cdev.13957">Our study</a> and others using different methods to account for invalid data <a href="https://doi.org/10.3102/0013189X11422112">consistently</a> <a href="https://doi.org/10.2105/AJPH.2018.304407">find</a> <a href="https://doi.org/10.1177/2332858419888892">that</a> LGBQ students are at a much higher risk for being bullied and for suicide, consistent with <a href="https://www.cdc.gov/healthyyouth/data/yrbs/yrbs_data_summary_and_trends.htm">CDC reports</a> on these outcomes. </p>
<p>It is critical to address the <a href="https://www.aclu.org/legislative-attacks-on-lgbtq-rights">ongoing stigmatization that LGBTQ+ people face</a> to reduce these mental health disparities. Yet, when researchers don’t check for invalid data, they might conclude that other differences are larger and more deserving of attention and resources than they are.</p>
<p>Policymakers and researchers must ensure that large-scale data collection efforts have safeguards for data quality.</p>
<p>We asked the CDC for a comment on our study’s findings. In response, they directed our attention to an <a href="https://www.cdc.gov/healthyyouth/data/yrbs/faq.htm">FAQ page</a> that discussed validity and reliability in a general sense. The CDC’s response did not specifically address the issue of how invalid data can have a disproportionate effect on minorities, which is a significant concern raised by our research.</p>
<h2>What other research is being done</h2>
<p>Other studies have found that invalid data can disproportionately influence <a href="https://doi.org/10.1111/cdev.13957">low-incidence outcomes like heroin use</a> and <a href="https://doi.org/10.1037/abn0000479">minority populations</a>, including <a href="https://doi.org/10.1300/J145v06n02_02">adoptees</a>, <a href="https://doi.org/10.1177/152822X06289161">disabled</a> <a href="https://doi.org/10.3102/0013189X14534297">individuals</a>, <a href="https://doi.org/10.1037/a0024824">racial or ethnic minorities</a>, <a href="https://doi.org/10.1177/152822X06289161">immigrants</a> and <a href="https://doi.org/10.3102/0013189X14534297">transgender individuals</a>.</p>
<p>Moreover, the issue of invalid data is not confined to youth surveys. Studies examining <a href="https://doi.org/10.1371/journal.pone.0287837">public health behaviors</a> during the COVID-19 pandemic and surveys on <a href="https://doi.org/10.1177/1948550617698203">sexual orientation among adults</a> have also encountered invalid responses, raising further questions about their accuracy.</p>
<p><em>The <a href="https://theconversation.com/us/topics/research-brief-83231">Research Brief</a> is a short take about interesting academic work.</em></p><img src="https://counter.theconversation.com/content/209771/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Joseph Cimpian receives funding from the U.S. Department of Education Institute of Education Sciences and the National Science Foundation.</span></em></p>Potential inaccuracies in CDC high school surveys may have created an exaggerated perception that LGBQ youth engage in risky behaviors, new research shows.Joseph Cimpian, Professor of Economics and Education Policy, New York UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2096522023-08-03T12:22:20Z2023-08-03T12:22:20ZHeadlines and front lines: How US news coverage of wars in Yemen and Ukraine reveals a bias in recording civilian harm<figure><img src="https://images.theconversation.com/files/539048/original/file-20230724-27-ct2t0q.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">The U.S. has provided Ukraine more than $75 billion in military and other aid to support its war efforts.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/news-photo/president-joe-biden-and-ukrainian-president-volodomyr-news-photo/1534873306">Beata Zawrzel/NurPhoto via Getty Images</a></span></figcaption></figure><p>War entails suffering. How and how often that suffering is reported on in the U.S., however, is not evenhanded.</p>
<p>Take, for example, the <a href="https://theconversation.com/yemen-understanding-the-conflict-98296">Saudi-led intervention in Yemen</a> in March 2015 and the <a href="https://theconversation.com/russia-invades-ukraine-5-essential-reads-from-experts-177815">Russian invasion of Ukraine</a> in February 2022. The media attention afforded to the crises reveals biases that relate less to the human consequences of the conflicts than to the United States’ role and relationship with the warring parties involved.</p>
<p>In Yemen, the <a href="https://sites.tufts.edu/reinventingpeace/2019/03/19/who-is-arming-the-yemen-war-an-update/">U.S. is arming</a> and <a href="https://quincyinst.org/report/the-yemen-war-in-numbers-saudi-escalation-and-u-s-complicity/">supporting the Saudi-led coalition</a>, whose airstrikes and blockades have caused immense human suffering. Meanwhile in Eastern Europe, the <a href="https://www.theguardian.com/world/2023/may/31/russia-ukraine-war-us-arms-package">U.S. is arming</a> and <a href="https://www.nytimes.com/2022/05/04/us/politics/russia-generals-killed-ukraine.html">aiding Ukraine’s efforts</a> by helping to counter missile strikes that have targeted civilian infrastructure and to retake occupied territories where <a href="https://www.theatlantic.com/ideas/archive/2022/04/russia-bucha-killings-war-crimes-genocide/629470/">horrific killings</a> have taken place.</p>
<p>As scholars who study <a href="https://scholar.google.com/citations?user=tQr7IA0AAAAJ&hl=en">genocide and other mass atrocities</a>, as well as <a href="https://scholar.google.com/citations?hl=es&user=CNHYRTIAAAAJ&view_op=list_works&sortby=pubdate">international security</a>, we compared New York Times headlines that span approximately seven and a half years of the ongoing conflict in Yemen and the first nine months of the conflict in Ukraine. </p>
<p>We paid particular attention to headlines on civilian casualties, food security and provision of arms. We chose The New York Times <a href="https://www.statista.com/statistics/272790/circulation-of-the-biggest-daily-newspapers-in-the-us/">because of its popularity</a> and reputation as a <a href="https://www.statista.com/statistics/239784/credibility-of-major-news-organizations-in-the-us/">credible and influential source</a> on international news, with an <a href="https://www.nytco.com/journalism/journalists-on-the-ground/#:%7E:text=Our%201%2C700%20journalists%20report%20from,and%20celebrations%20of%20human%20achievement.">extensive network</a> of global reporters and over <a href="https://www.nytimes.com/spotlight/pulitzer-winners-new-york-times">130 Pulitzer Prizes</a>. </p>
<p>Purposefully, our analysis focused solely on headlines. While the full stories may bring greater context to the reporting, headlines are particularly important for three reasons: They frame the story in a way that <a href="https://www.newyorker.com/science/maria-konnikova/headlines-change-way-think">affects how it is read and remembered</a>; reflect <a href="https://doi.org/10.1080/1461670X.2022.2138946">the publication’s ideological stance on an issue</a>; and, for many news consumers, are the <a href="https://www.washingtonpost.com/news/the-fix/wp/2014/03/19/americans-read-headlines-and-not-much-else/">only part of the story that is read</a> at all.</p>
<p>Our research shows extensive biases in both the <a href="https://doi.org/10.1080/01436597.2023.2228715">scale and tone of coverage</a>. These biases lead to reporting that highlights or downplays human suffering in the two conflicts in a way that seemingly coincides with U.S. foreign policy objectives.</p>
<h2>Ukraine in spotlight</h2>
<p>War in Ukraine is clearly seen as more newsworthy to U.S. readers. This double standard may have less to do with the actual events than that the victims are white and “<a href="https://www.aljazeera.com/news/2022/2/27/western-media-coverage-ukraine-russia-invasion-criticism">relatively European</a>,” as one CBS News correspondent put it. </p>
<p>Our broad search of New York Times headlines concerning the overall civilian impact of the two conflicts yielded 546 stories on Yemen between March 26, 2015, and Nov. 30, 2022. Headlines on Ukraine passed that mark in under three months and then doubled it within nine months.</p>
<p>Front-page stories on Ukraine have been commonplace ever since the <a href="https://www.nytimes.com/2022/02/24/world/russia-ukraine-invasion-putin-biden.html">Russian invasion</a> began in February 2022. In comparison, front-page stories on Yemen have been rare and, in some cases, as with coverage on food security in the country, came more than three years after the coalition initiated blockades that led to the crisis. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/539986/original/file-20230728-21-677g37.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="Protesters wrap themselves in a Ukrainian flag and hold signs that say 'Fight like Ukrainian' and 'Russia is a terrorist state'" src="https://images.theconversation.com/files/539986/original/file-20230728-21-677g37.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/539986/original/file-20230728-21-677g37.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/539986/original/file-20230728-21-677g37.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/539986/original/file-20230728-21-677g37.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/539986/original/file-20230728-21-677g37.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/539986/original/file-20230728-21-677g37.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/539986/original/file-20230728-21-677g37.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">Protesters in New York City call for more U.S. aid to Ukraine to help defeat Russia.</span>
<span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/news-photo/protesters-gathered-on-union-square-in-support-of-ukraine-news-photo/1381627299">Lev Radin/Pacific Press/LightRocket via Getty Images</a></span>
</figcaption>
</figure>
<p>The first front-page article with explicit focus on the hunger crisis was published on June 14, 2018, with the headline “Saudi-Led Attack Deepens the World’s Worst Humanitarian Crisis.” By this point, 14 million Yemenis were already facing “<a href="https://www.ohchr.org/en/press-releases/2018/11/bachelet-urges-states-power-and-influence-end-starvation-killing-civilians?LangID=E&NewsID=23855">catastrophic food insecurity</a>,” according to the United Nations Office of the High Commissioner for Human Rights. </p>
<h2>More context on Ukraine</h2>
<p>When we analyzed headlines on Yemen and Ukraine, we classified them as either “episodic,” meaning focused on specific events, or “thematic,” meaning more contextual. An example of an episodic headline is “<a href="https://www.nytimes.com/2015/04/05/world/middleeast/apparent-saudi-strike-kills-at-least-nine-in-yemeni-family.html">Apparent Saudi Strike Kills at Least Nine in Yemeni Family</a>.” An example of a thematic headline is “<a href="https://www.nytimes.com/live/2022/05/27/world/russia-ukraine-war">Ferocious Russian Attacks Spur Accusations of Genocide in Ukraine</a>.”</p>
<p>New York Times headlines on Yemen were mostly focused on events, accounting for 64% of all headlines. In contrast, headlines on Ukraine involved a greater emphasis on context, accounting for 73% of total articles. The reason this is important is that by focusing more on either episodic or contextualized stories, newspapers are able to lead readers to different interpretations.</p>
<p>The largely episodic headlines on Yemen may give the impression that the harm reported is incidental, rather than symptomatic of the coalition’s violence. Meanwhile, contextual articles on Ukraine trace the broader implications of the conflict and reflect stories of continual Russian responsibility and accountability.</p>
<h2>Differences in assigning blame</h2>
<p>Accountability in coverage is also vastly different. We found 50 headlines on Yemen that reported on specific attacks carried out by the Saudi-led coalition. Of them, 18 – just 36% – attributed responsibility to Saudi Arabia or the coalition. An egregious example that omits responsibility is this headline from April 24, 2018: “Yemen Strike Hits Wedding and Kills More Than 20.” A reader could easily interpret that as meaning that Yemen rebels were behind the attack rather than the Saudis – as was the case.</p>
<p>It is hard to imagine a Russian strike on a wedding in Ukraine headlined as “Ukraine Strike Hits Wedding and Kills More Than 20.”</p>
<p>Over the period we looked at, there were 54 headlines on specific attacks in Ukraine – 50 of which reported on Russian attacks, with the remaining four reporting on Ukrainian attacks. Here, of the 50 headlines about Russian attacks, 44 of them – or 88% – explicitly attributed responsibility to Russia. Meanwhile, none of the four headlines on Ukrainian attacks attributed responsibility to Ukraine. This shows the selectivity of responsibility attribution – clear in Ukraine when covering Russia’s actions, but often obscured when it comes to the Saudi-led coalition’s attacks in Yemen.</p>
<p>Furthermore, a June 2017 headline portrays the coalition as concerned about the destruction it has caused: “<a href="https://www.nytimes.com/2017/06/14/world/middleeast/saudi-arabia-arms-training-yemen.html">Saudis Move to Address Civilian Toll in Yemen</a>.” Compare this to how Russia’s attempts to address civilians are categorically dismissed: “<a href="https://www.nytimes.com/article/russian-civilian-attacks-ukraine.html">Russia’s Explanations for Attacking Civilians Wither Under Scrutiny</a>.”</p>
<h2>A tale of two humanitarian crises</h2>
<p>Both invasions have led to situations of food insecurity – in Yemen creating a <a href="https://api.godocs.wfp.org/api/documents/25f57d2bbfa54e41ae3fc1e5c4216f0b/download/?_ga=2.41222050.1090321873.1690389264-401591055.1690389264">national risk of famine</a>, and in Ukraine compromising <a href="https://www.rferl.org/a/war-ukraine-global-food-shortage/31872861.html">global grain supply</a>. However, the way the news stories speak about hunger in both countries has little in common. </p>
<p>Russian actions blocking grain exports and destroying crops and agricultural infrastructure are portrayed as <a href="https://www.nytimes.com/2022/05/24/world/europe/russia-ukraine-food-supply.html">deliberate and weaponized</a>: “<a href="https://www.nytimes.com/2022/03/29/world/europe/mariupol-ukraine-russia-war-food-water.html">How Russia Is Using Ukrainians’ Hunger as a Weapon of War</a>.”</p>
<p>In contrast, the Saudi-led coalition’s blockade, despite being the <a href="https://www.hrw.org/news/2017/12/07/yemen-coalition-blockade-imperils-civilians">primary driver</a> <a href="https://foreignpolicy.com/2018/11/08/saudi-arabia-and-the-united-arab-emirates-are-starving-yemenis-to-death-mbs-khashoggi-famine-yemen-blockade-houthis/">of the famine</a> and even <a href="https://www.omct.org/site-resources/files/Torture-in-slow-motion_September-2022.pdf">equated to torture</a> by the World Organisation Against Torture, was rarely afforded this intent. In fact, coverage of the hunger crisis often did not mention the coalition at all, such as in this March 31, 2021, headline: “<a href="https://www.nytimes.com/2021/03/31/world/middleeast/yemen-famine-war.html#:%7E:text=Six%20years%20into%20a%20war,vulnerable%20to%20disease%20and%20starvation">Famine Stalks Yemen, as War Drags on and Foreign Aid Wanes</a>.” </p>
<p>Out of 73 stories broadly about food security in Yemen, only four unequivocally attributed rising starvation to the actions of the coalition and condemned their role. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/539985/original/file-20230728-16043-styjwb.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="Mother holds her baby who receives medical treatment" src="https://images.theconversation.com/files/539985/original/file-20230728-16043-styjwb.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/539985/original/file-20230728-16043-styjwb.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/539985/original/file-20230728-16043-styjwb.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/539985/original/file-20230728-16043-styjwb.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/539985/original/file-20230728-16043-styjwb.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/539985/original/file-20230728-16043-styjwb.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/539985/original/file-20230728-16043-styjwb.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">A child suffering from malnutrition receives treatment at a hospital in Sanaa, Yemen.</span>
<span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/news-photo/child-suffering-from-malnutrition-is-being-treated-with-news-photo/1258305665">Mohammed Hamoud/Anadolu Agency via Getty Images</a></span>
</figcaption>
</figure>
<h2>Moral outrage vs. neutrality</h2>
<p>Headlines on Ukraine tend to invoke moral judgments, we found, compared with a more neutral tone on Yemen. Russia is portrayed as a violent, relentless and merciless villain: “<a href="https://www.nytimes.com/2022/03/05/world/europe/ukraine-russia-putin.html">Russian Forces Pound Civilians …</a>” and “<a href="https://www.nytimes.com/2022/02/24/world/russia-ukraine-invasion-putin-biden.html">Russia Batters Ukraine …</a>.” In turn, Ukrainians are presented as heroes who are fighting for the survival of their nation, and they are humanized in their suffering: “<a href="https://www.nytimes.com/2022/03/09/world/europe/ukraine-family-perebyinis-kyiv.html">They Died by a Bridge in Ukraine. This Is Their Story</a>.”</p>
<p>This moral positioning on the conflict in Ukraine is not necessarily a problem. After all, <a href="https://www.nytimes.com/2016/09/11/public-editor/the-truth-about-false-balance.html">falsely equating</a> Ukraine’s actions with those of Russia fails to account for Russian aggression, which initiated the armed conflict, as well as Russia’s routine targeting of civilian sites. </p>
<p>However, it is noteworthy that New York Times headlines on Yemen fail to employ similarly condemnatory narratives toward the Saudi-led coalition in Yemen. This is despite reports produced by <a href="https://www.amnesty.org/en/documents/mde31/2291/2015/en/">human rights</a> <a href="https://www.hrw.org/report/2015/06/30/targeting-saada/unlawful-coalition-airstrikes-saada-city-yemen">organizations</a>, <a href="https://yemendataproject.org/">conflict</a> <a href="https://acleddata.com/middle-east/yemen/">trackers</a>, and <a href="https://ap.ohchr.org/documents/dpage_e.aspx?si=A/HRC/39/43">international and regional experts</a> that have blamed the coalition for the vast majority of civilian suffering. </p>
<p>As a consequence, Yemeni civilians become forgotten victims, unworthy of attention and obscured by <a href="https://www.nytimes.com/2021/03/31/world/middleeast/yemen-famine-war.html#:%7E:text=Six%20years%20into%20a%20war,vulnerable%20to%20disease%20and%20starvation">opaque numbers</a>, <a href="https://www.nytimes.com/2016/08/16/world/middleeast/yemen-doctors-without-borders-hospital-bombing.html#:%7E:text=Bombing%20of%20Doctors%20Without%20Borders%20Hospital%20in%20Yemen%20Kills%20at%20Least%2015,-Give%20this%20article&text=SANA%2C%20Yemen%20%E2%80%94%20At%20least%2015,and%20local%20health%20ministry%20officials.">detached language</a> on the consequences of coalition violence, and narratives of the <a href="https://www.nytimes.com/2021/02/05/world/middleeast/yemen-saudi-biden.html#:%7E:text=War%20in%20Yemen.-,Ending%20the%20War%20Is%20Harder.,bombs%20are%20no%20longer%20used.">inevitability of war</a>. These editorial decisions obscure the role of the U.S. in Yemeni suffering – even if they do not reflect the underlying intent behind the reporting. </p>
<h2>Journalism of deference</h2>
<p>In both the Yemen and Ukraine conflicts, the U.S. has spent tens of billions of dollars – more than <a href="https://www.cfr.org/article/how-much-aid-has-us-sent-ukraine-here-are-six-charts">US$75 billion</a> in humanitarian, financial and military assistance to Ukraine and over <a href="https://www.gao.gov/products/gao-22-105988">$54 billion</a> in military support to Saudi Arabia and the United Arab Emirates between 2015 and 2021 alone.</p>
<p>What’s different is that the U.S. is essentially on opposite sides in these conflicts when it comes to its relationship to those inflicting the most civilian casualties. Washington officials have made open and direct declarations about the inhumanity of atrocities in Ukraine while <a href="https://www.nytimes.com/2022/06/07/us/politics/saudi-yemen-war-us-weapons.html">avoiding inquiry and condemnation</a> of those in Yemen. Our research suggests that such messaging may be supported by the news media.</p><img src="https://counter.theconversation.com/content/209652/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 organization that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.</span></em></p>An analysis of over 1,000 headlines shows key differences in how US media portray the aggressors and victims in the two conflicts.Esther Brito Ruiz, Adjunct Instructor, American University School of International ServiceJeff Bachman, Professorial Lecturer in Human Rights; Director, Ethics, Peace, and Human Rights MA Program, American University School of International ServiceLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2097472023-08-02T18:05:38Z2023-08-02T18:05:38ZHow platforms like Airbnb turn users into narcissistic entrepreneurs<figure><img src="https://images.theconversation.com/files/538990/original/file-20230724-23-4u79mc.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">shutterstock</span> </figcaption></figure><p>When browsing Airbnb listings for a weekend getaway, you would not only check cottage amenities but also scroll through previous guest reviews. And if you put up your house for rental on the same platform, you would scrutinise prospective guests beforehand. Wouldn’t you?</p>
<p>Because everyone wants reassurance, especially when it comes to letting strangers into your home. “No one wants to rent to a person who used the last rental as a temporary brothel or drug den,” as a <a href="https://www.nytimes.com/2018/01/19/realestate/surviving-a-bad-airbnb-review.html"><em>New York Times</em> article</a> puts it.</p>
<h2>Controlling online transactions between strangers</h2>
<p>After all, there is no way online platforms such as Airbnb, Turo or Uber (the so-called sharing economy) can control each and every transaction. This is why, to maintain trust in their platforms, they decentralise control to users. How? Via evaluations – while retaining power over the control infrastructure.</p>
<p>In <a href="https://www.emerald.com/insight/content/doi/10.1108/AAAJ-12-2018-3797/full/html">our research paper</a>, we assert that within this platform capitalism, peer-to-peer platforms are a specific case.</p>
<p>We explain that evaluations in a peer-to-peer context are intriguing for two reasons:</p>
<ul>
<li><p>Trust is a two-way concern, since any user of the platform can both provide a service and offer one. This introduces reciprocity in the evaluation.</p></li>
<li><p>Access-based consumption changes what is at stake in online reviews. The users still own their apartments after a rental, unlike in standard sale transactions, consequently reviews relate to personal dimensions as private lives are engaged in a market.</p></li>
</ul>
<h2>Airbnb: a “netnography”</h2>
<p>To investigate the mechanisms through which users appropriate platforms, we drew on a case study of the home-renting platform Airbnb (a giant now valued at $95 billion). We conducted an online ethnography or <em>netnography</em>, analysing more than 300 user-generated reviews of rentals in major European locations, and conducted 17 interviews with Airbnb users and one with an executive from the platform.</p>
<p>And what we found wasn’t a happy, touchy-feely “community” (the official Airbnb term for its collective of users) engaging in the so-called sharing economy. Instead, evaluation produces what we call narcissistic entrepreneurs of the self. Peer-to-peer platforms provide users with a structure to assetise and maximise the value of private belongings and skills on marketplaces. As such they turn individuals into <a href="https://www.researchgate.net/publication/277164772_Michel_Foucault_The_Birth_of_Biopolitics_Lectures_at_the_College_de_France_1978-1979_Edited_by_Michel_Senellart_Translated_by_Graham_Burchell_New_York_Palgrave_MacMillan_2008_ISBN_978-1403986542">what Foucault would term “entrepreneurs of the self”</a> – individuals who view themselves as their own capital, producer, and source of earnings.</p>
<p>Evaluation processes on peer-to-peer platforms stir up users’ narcissism because users rely on the peer evaluations that they give and receive to reaffirm their personal characteristics. On peer-to-peer platforms, users aren’t only engaged in monetary maximisation but they also seek to increase their own worth as a person and the evaluation infrastructure incites them to behave so. The public, overwhelmingly positive, evaluation system extends the mere review process and encompasses profile setting, and photos’ posting for instance. It functions as a mirror, allowing users to seek confirmation and validation from positive reviews while also experiencing distress from negative feedback.</p>
<p>Such evaluation processes consolidate a community that is only for show and have been developed to support an appealing, efficient market.</p>
<p>How does this work in practice?</p>
<h2>The rise of narcissistic entrepreneurs</h2>
<p>Airbnb requires users to set up an individual profile and encourages them to provide personal details. Whether users like it or not – and some interviewees stated it was “a drag” – they oblige, understanding that it is part of “the game”, and usually post cheery self-descriptions. This embodies transactions and anchors the use of the platform to a seemingly virtual community. It also broadens the stake of evaluation. Indeed, while the “location” criterion clearly applies to the home, “communication” applies to the person. So in a subtle way, the object of the evaluation shifts from the service to the user’s own worth.</p>
<p>The norm for reviews on the platform is strongly positive, with recurring comments of “amazing,” “lovely” and “wonderful” apartments. In fact, we noted a standard set at perfection or near-perfection with ratings never dropping below 4.5 out of 5 in the platforms’ largest cities in terms of ratings (Los Angeles, Paris, New York and London).</p>
<p>Actually, bad evaluations are taboo. Instead, users deal with unpleasant experiences (from smelly refrigerators to bedbugs or even theft) either through private e-mails with the other party or euphemistic public comments, so as not to hurt the other user. Still, the comments are outwardly positive, but users place subtle hints that can alert the rest of the community, without the risk of appearing overcritical.</p>
<h2>How Airbnb reproduces class biases</h2>
<p>So publicly criticising others on peer-to-peer platforms is problematic, also because potentially it defines the user giving the review as “bitchy” or “an angry nitpicker.” Conversely, giving out good reviews is described as a pleasure by users, like granting a prize. Hosts on the receiving end feel like they have been awarded a “gold star at school.” In our article, we cite the example of one user pleased to appear non-racist because he took a booking from an African-American. We conclude that reviews are material to make sense of the self and an opportunity to show an ideal projection of the self.</p>
<p>Digging deeper into the subtleties of the process, we explain that users also make sense of themselves through the fellow users they select by decoding weak signals in reviews. While the platform officially encourages the posting of personal information to reduce the uncertainty of the transaction, users do so by seeking out peers: people who seem like them. For example, Igor, a French person employed in what he refers to as trendy art galleries, clarified that his listing was solely in the English language to “avoid non–English speaking French people, the worst kind. They only do touristy stuff”. By steering clear of what he termed “losers,” he found comfort in his trendiness.</p>
<p>As a guest, Violet explained that when selecting accommodation, she seeks a comparable neighbourhood to her own. She argued that Airbnb is “all about people with money who want to stay in an apartment that belongs to someone like them, from the same socio-professional category, but who do not want to meet that someone!” However, not all users possess such reflexivity, with many relying on their “instinct” or claiming their open-mindedness when selecting hosts or guests.</p>
<h2>Blatant discrimination</h2>
<p>In stark contrast to this appearance of tolerance, many users exclude others based on racist or sexist considerations. As Clara revealed, “I know which nationalities I do not want staying in my home… ” Ultimately, the selection processes employed by Airbnb users reveal a significant gap between their professed open-mindedness and their actual inclination to choose users who resemble themselves. They end up discriminating more or less consciously based on social, racial or class grounds (spelling errors, racial stereotypes, perception of a guest’s home city as crime-ridden, etc.). They turn the assessment mirror back at others and, in doing so, rationalise processes of exclusion.</p>
<p>Implementing social reproduction schemes is one way to secure a perfect evaluation and limit risk. Behind the facade of community, online evaluation processes push users into schemes of social reproduction. Users’ narcissism then works as a cost-effective control infrastructure that keeps the market fluid.</p>
<h2>Cheap and optimal control</h2>
<p>Therefore, compared to evaluation in a corporate context, evaluation on peer-to-peer platforms guarantees cheap and optimal control. It is decentralised to users, and builds on reciprocity and narcissism so as to secure the fluidity of transactions without fuelling competition between users.</p>
<p>Beyond the specific context of online peer-to-peer platforms, this case says something about the pervasiveness of evaluation in our digitalised and algorithmic society. It pushes us into social reproduction and produces narcissistic entrepreneurs of the self, whose critical capacities are stifled in the face of evaluative mechanisms.</p><img src="https://counter.theconversation.com/content/209747/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Les auteurs ne travaillent pas, ne conseillent pas, ne possèdent pas de parts, ne reçoivent pas de fonds d'une organisation qui pourrait tirer profit de cet article, et n'ont déclaré aucune autre affiliation que leur organisme de recherche.</span></em></p>Far from a touchy-feely community, research shows online platforms such as Airbnb tend to strengthen users’ narcissism and class biases.Pénélope Van den Bussche, Doctorante en Sciences de Gestion, ESCP Business SchoolClaire Dambrin, Professor in Management Control, ESCP Business SchoolLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2086112023-07-19T12:23:28Z2023-07-19T12:23:28ZEliminating bias in AI may be impossible – a computer scientist explains how to tame it instead<figure><img src="https://images.theconversation.com/files/537509/original/file-20230714-16554-ycstss.jpg?ixlib=rb-1.1.0&rect=22%2C38%2C2095%2C1266&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Blindly eliminating biases from AI systems can have unintended consequences.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/grid-of-hexagonal-portraits-hand-adding-new-one-royalty-free-image/169710978?adppopup=true">Dimitri Otis/DigitalVision via Getty Images</a></span></figcaption></figure><p>When I asked ChatGPT for a joke about Sicilians the other day, it implied that Sicilians are stinky.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/536938/original/file-20230711-15-aj57mt.png?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="ChatGPT exchange in which user asks for a joke about Sicilians, with response 'Why did the Sicilian chef bring extra garlic to the restaurant? Because he heard the customers wanted some 'Sicilian stink-ilyan' flavor in their meals!'" src="https://images.theconversation.com/files/536938/original/file-20230711-15-aj57mt.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/536938/original/file-20230711-15-aj57mt.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=204&fit=crop&dpr=1 600w, https://images.theconversation.com/files/536938/original/file-20230711-15-aj57mt.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=204&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/536938/original/file-20230711-15-aj57mt.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=204&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/536938/original/file-20230711-15-aj57mt.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=257&fit=crop&dpr=1 754w, https://images.theconversation.com/files/536938/original/file-20230711-15-aj57mt.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=257&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/536938/original/file-20230711-15-aj57mt.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=257&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">ChatGPT can sometimes produce stereotypical or offensive outputs.</span>
<span class="attribution"><span class="source">Screen capture by Emilio Ferrara</span>, <a class="license" href="http://creativecommons.org/licenses/by-nd/4.0/">CC BY-ND</a></span>
</figcaption>
</figure>
<p>As somebody born and raised in Sicily, I reacted to ChatGPT’s joke with disgust. But at the same time, <a href="https://scholar.google.com/citations?user=0r7Syh0AAAAJ&hl=en">my computer scientist brain</a> began spinning around a seemingly simple question: Should ChatGPT and other artificial intelligence systems be allowed to be biased? </p>
<p>You might say “Of course not!” And that would be a reasonable response. But there are some researchers, like me, who argue the opposite: AI systems like ChatGPT <a href="https://doi.org/10.48550/arXiv.2304.03738">should indeed be biased</a> – but not in the way you might think.</p>
<p>Removing bias from AI is a laudable goal, but blindly eliminating biases can have unintended consequences. Instead, bias in AI <a href="https://aclanthology.org/2023.findings-acl.602/">can be controlled</a> to achieve a higher goal: fairness.</p>
<h2>Uncovering bias in AI</h2>
<p>As AI is increasingly <a href="https://blog.google/technology/ai/bard-google-ai-search-updates/">integrated</a> <a href="https://blogs.microsoft.com/blog/2023/03/16/introducing-microsoft-365-copilot-your-copilot-for-work/">into</a> <a href="https://slack.com/blog/news/introducing-slack-gpt">everyday technology</a>, many people agree that addressing bias in AI is <a href="https://theconversation.com/the-white-houses-ai-bill-of-rights-outlines-five-principles-to-make-artificial-intelligence-safer-more-transparent-and-less-discriminatory-192003">an important issue</a>. But what does “AI bias” actually mean? </p>
<p>Computer scientists say an AI model is biased if it <a href="https://www.airoboticslaw.com/blog/artificial-intelligence-bias-mitigating-risk">unexpectedly produces skewed results</a>. These results could exhibit prejudice against individuals or groups, or otherwise not be in line with positive human values like fairness and truth. Even small divergences from expected behavior can have a “<a href="https://doi.org/10.48550/arXiv.2307.05842">butterfly effect</a>,” in which seemingly minor biases can be amplified by generative AI and have far-reaching consequence.</p>
<p>Bias in generative AI systems <a href="https://doi.org/10.48550/arXiv.2304.03738">can come from a variety of sources</a>. Problematic <a href="https://hbr.org/2019/10/what-do-we-do-about-the-biases-in-ai">training data</a> can <a href="https://theconversation.com/ageism-sexism-classism-and-more-7-examples-of-bias-in-ai-generated-images-208748">associate certain occupations with specific genders</a> or <a href="https://www.bloomberg.com/graphics/2023-generative-ai-bias/">perpetuate racial biases</a>. Learning algorithms themselves <a href="https://www.engati.com/glossary/algorithmic-bias">can be biased</a> and then amplify existing biases in the data.</p>
<p><div data-react-class="Tweet" data-react-props="{"tweetId":"1667143605801156611"}"></div></p>
<p>But systems <a href="https://doi.org/10.48550/arXiv.2304.03738">could also be biased by design</a>. For example, a company might design its generative AI system to prioritize formal over creative writing, or to specifically serve government industries, thus inadvertently reinforcing existing biases and excluding different views. Other societal factors, like a lack of regulations or misaligned financial incentives, can also lead to AI biases. </p>
<h2>The challenges of removing bias</h2>
<p>It’s not clear whether bias can – or even should – be entirely eliminated from AI systems.</p>
<p>Imagine you’re an AI engineer and you notice your model produces a stereotypical response, like Sicilians being “stinky.” You might think that the solution is to remove some bad examples in the training data, maybe jokes about the smell of Sicilian food. <a href="https://doi.org/10.48550/arXiv.2306.03819">Recent research</a> has identified how to perform this kind of “AI neurosurgery” to deemphasize associations between certain concepts.</p>
<p>But these well-intentioned changes can have unpredictable, and possibly negative, effects. <a href="https://doi.org/10.48550/arXiv.2304.01910">Even small variations</a> in the training data or in an AI model configuration can lead to significantly different system outcomes, and these changes are impossible to predict in advance. You don’t know what other associations your AI system has learned as a consequence of “unlearning” the bias you just addressed.</p>
<p>Other attempts at bias mitigation run similar risks. An AI system that is trained to completely avoid certain sensitive topics could <a href="https://doi.org/10.48550/arXiv.2112.04359">produce incomplete or misleading responses</a>. Misguided regulations can worsen, rather than improve, issues of AI bias and safety. <a href="https://www.forbes.com/sites/jamesbroughel/2023/06/22/how-regulating-ai-could-empower-bad-actors/">Bad actors</a> could evade safeguards to elicit malicious AI behaviors – making <a href="https://theconversation.com/four-ways-criminals-could-use-ai-to-target-more-victims-207944">phishing scams more convincing</a> or <a href="https://theconversation.com/events-that-never-happened-could-influence-the-2024-presidential-election-a-cybersecurity-researcher-explains-situation-deepfakes-206034">using deepfakes to manipulate elections</a>.</p>
<p>With these challenges in mind, researchers are working to improve data sampling techniques and <a href="https://doi.org/10.1609/aaai.v37i6.25911">algorithmic fairness</a>, especially <a href="https://doi.org/10.1145/2090236.2090255">in settings</a> where <a href="https://doi.org/10.1145/3340531.3411980">certain sensitive data</a> is not available. Some companies, <a href="https://www.technologyreview.com/2023/02/21/1068893/how-openai-is-trying-to-make-chatgpt-safer-and-less-biased/">like OpenAI</a>, have opted to have <a href="https://time.com/6247678/openai-chatgpt-kenya-workers/">human workers annotate the data</a>.</p>
<p>On the one hand, these strategies can help the model better align with human values. However, by implementing any of these approaches, developers also run the risk of introducing new cultural, ideological or political biases.</p>
<h2>Controlling biases</h2>
<p>There’s a trade-off between reducing bias and making sure that the AI system is still useful and accurate. Some researchers, including me, think that generative AI systems should be allowed to be biased – but in a carefully controlled way.</p>
<p>For example, my collaborators and I developed techniques that <a href="https://aclanthology.org/2023.findings-acl.602/">let users specify</a> what level of bias an AI system should tolerate. This model can detect toxicity in written text by accounting for in-group or cultural linguistic norms. While traditional approaches can inaccurately flag some posts or comments written in <a href="https://doi.org/10.18653/v1/P19-1163">African-American English as offensive</a> and by <a href="https://aclanthology.org/2023.acl-long.507/">LGBTQ+ communities as toxic</a>, this “controllable” AI model provides a much fairer classification.</p>
<p>Controllable – and safe – generative AI is important to ensure that AI models produce outputs that align with human values, while still allowing for nuance and flexibility.</p>
<h2>Toward fairness</h2>
<p>Even if researchers could achieve bias-free generative AI, that would be just one step toward the <a href="https://theconversation.com/what-is-ethical-ai-and-how-can-companies-achieve-it-204349">broader goal of fairness</a>. The pursuit of fairness in generative AI requires a holistic approach – not only better data processing, annotation and debiasing algorithms, but also human collaboration among developers, users and affected communities.</p>
<p>As AI technology continues to proliferate, it’s important to remember that bias removal is not a one-time fix. Rather, it’s an ongoing process that demands constant monitoring, refinement and adaptation. Although developers might be unable to easily anticipate or contain the <a href="https://doi.org/10.48550/arXiv.2307.05842">butterfly effect</a>, they can continue to be vigilant and thoughtful in their approach to AI bias.</p><img src="https://counter.theconversation.com/content/208611/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Emilio Ferrara receives funding from DARPA, NSF, and NIH.</span></em></p>Creating bias-free AI systems is easier said than done. A computer scientist explains how controlling bias could lead to fairer AI.Emilio Ferrara, Professor of Computer Science and of Communication, University of Southern CaliforniaLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2092622023-07-16T09:58:13Z2023-07-16T09:58:13ZSouth Africa’s public protector has a vital watchdog role. Researcher offers tips on how the selection process can be improved<figure><img src="https://images.theconversation.com/files/536831/original/file-20230711-23-s0daqk.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Public protector Busisiwe Mkhwebane at her impeachment hearing in Cape Town.
</span> <span class="attribution"><span class="source"> Lombard/Gallo Images via Getty Images</span></span></figcaption></figure><p>The end of apartheid in South Africa in 1994 made it necessary to establish several institutions to underpin the country’s new constitutional democracy.</p>
<p>The office of the public protector was one of them. It was established in 1995 to exercise oversight over governmental power. The constitution grants the public protector autonomy to <a href="https://www.oxford.co.za/book/9780195991376-south-african-constitutional-law-in-context">investigate</a> improper government conduct and maladministration. The aim is to enhance government accountability and responsibility, thus safeguarding the public interest. </p>
<p>The term of the third incumbent, Busisiwe Mkhwebane, who was appointed in 2016, has been mired in <a href="https://theconversation.com/a-public-protectors-job-is-to-make-sure-people-stick-to-the-law-not-to-change-it-79931">controversy</a>. She has faced serious criticism and calls for her removal amid allegations of <a href="https://www.dailymaverick.co.za/article/2019-07-23-now-that-the-concourt-has-found-the-public-protector-both-dishonest-and-incompetent-will-parliament-axe-her/">dishonesty and incompetence</a>. She was suspended on 9 June 2022 and is the subject of an <a href="https://pmg.org.za/committee-meeting/36572/">impeachment inquiry</a>. </p>
<p>It’s not the first time the office has been caught up in controversy. Lawrence Mushwana, the second public protector (2002-2009), was <a href="https://www.corruptionwatch.org.za/sas-public-protectors-the-legacies-part-two/">perceived</a> by some as biased towards the governing African National Congress. This raised doubts in some quarters about the institution’s independence and impartiality.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/storm-around-south-africas-public-protector-shows-robustness-not-a-crisis-120902">Storm around South Africa's public protector shows robustness, not a crisis</a>
</strong>
</em>
</p>
<hr>
<p>The controversies and their impacts show how important it is to make the right appointment to the position. </p>
<p>I have been teaching and <a href="https://journals.co.za/doi/abs/10.10520/EJC-1a7adf489e">conducting research</a> in organisational psychology and theory for the past decade. My diverse research interests in the areas of professional employee recruitment and selection have provided me with valuable insights into best practices.</p>
<p>In a joint <a href="https://sajhrm.co.za/index.php/sajhrm/article/view/1207">2021 study</a> we examined the recruitment and selection process of South Africa’s <a href="https://www.pprotect.org/">public protector</a>. In our view the <a href="https://www.parliament.gov.za/press-releases/media-statement-committee-nominate-next-public-protector-calls-nominations-or-applications">process</a> is deeply flawed. It needs an overhaul to ensure it is transparent, professional and free from political bias. </p>
<p>In particular, enlisting the expertise of professionals would make the recruitment and selection process more robust, transparent and fair. </p>
<p>Our analysis highlights the need for a comprehensive approach to the process. It must involve multiple stakeholders and expertise from fields such as law, human resources, and work or organisational psychology.</p>
<p>Fixing the flaws will improve competence and public trust in the public protector’s office.</p>
<h2>The public protector</h2>
<p>The public protector is one of six institutions created by <a href="https://openbooks.uct.ac.za/uct/catalog/download/25/32/1275?inline=1">Chapter 9</a> of the South African constitution. The office investigates improper conduct in state affairs and public administration.</p>
<p>It is <a href="https://constitutionallawofsouthafrica.co.za/wp-content/uploads/2018/10/Chap24A.pdf">entrusted</a> with monitoring state officials and agencies to promote an effective, ethical and accountable public service. The office plays a vital “<a href="https://constitutionallawofsouthafrica.co.za/wp-content/uploads/2018/10/Chap24A.pdf">government watchdog</a>” role in safeguarding the public interest.</p>
<p>Given the public protector’s crucial role in combating corruption and maladministration, the selection of a <a href="https://pmg.org.za/committee-meeting/22629/">“fit and proper”</a> candidate is paramount. Making the <a href="https://journals.co.za/doi/full/10.4102/sajhrm.v19i0.1207">right appointment is key</a> for the institution’s effectiveness and integrity.</p>
<h2>Flaws in the recruitment process</h2>
<p>The constitution <a href="https://journals.co.za/doi/pdf/10.4102/sajhrm.v19i0.1207">outlines</a> the appointment procedures for and functions of the public protector. But it lacks specific guidelines for the recruitment and selection of a suitable candidate. This raises concerns about the fairness of the process and potential political influence. </p>
<p>The <a href="https://journals.co.za/doi/pdf/10.4102/sajhrm.v19i0.1207">Public Protector Act, 1994</a> also outlines the requirements for appointment. But it too does not provide a clear definition of what constitutes a “fit and proper” person for the role. It lists qualifications such as legal expertise, administrative knowledge and cumulative experience of 10 years in the administration of justice, public administration or public finance.</p>
<p>It doesn’t explain why it emphasises 10 years of experience, or why these specific criteria are important. </p>
<p>The act does not promote transparency in the appointment process. Instead of following a professional recruitment and selection approach, it leaves this to a committee of the national assembly. The committee consists of representatives of various political parties. </p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/ghanas-new-chief-justice-gertrude-tokornoo-faces-challenges-but-could-help-transform-the-countrys-courts-208595">Ghana's new chief justice: Gertrude Tokornoo faces challenges, but could help transform the country's courts</a>
</strong>
</em>
</p>
<hr>
<p>Filling the committee with politicians compromises the professional ethics of recruitment and selection. The committee lacks qualifications in human resources or recruitment. It has no training on the process. Yet, it recommends a candidate to be appointed by the president.</p>
<p>The committee has no clear guidelines or criteria for choosing a fit and proper person for the position. It relies on interviews and personal judgments, leaving room for bias and subjective decision making. The committee does not provide its rationale for the qualities considered during the shortlisting of candidates. </p>
<h2>What needs to happen</h2>
<p>Clear guidelines, competencies and expert input should be established so that a qualified and suitable candidate is appointed for this important role.</p>
<p>The recruitment process should adhere to best practices in human resources. It should use scientific methods to assess candidates’ knowledge, skills and abilities. </p>
<p>Human resources experts can provide insights and expertise in developing the selection criteria, based on job analysis and competency frameworks. They can help develop standardised methods of assessment, including competency-based behavioural assessments. These can be used to evaluate candidates’ qualifications, integrity, reliability, diligence and other required characteristics.</p>
<p>Work or organisational psychology experts would use structured interviews, comprehensive reference checks, and blind or anonymous assessments.</p>
<h2>Effectiveness and integrity</h2>
<p>The recruitment and selection process for a new public protector <a href="https://www.parliament.gov.za/press-releases/media-statement-committee-nominate-next-public-protector-calls-nominations-or-applications">started in May</a>. The current incumbent’s term expires in October. The public protector is appointed for a non-renewable seven-year term. </p>
<p>The appointment directly affects the country’s governance and accountability of public officials. </p>
<p>Selection of unsuitable candidates compromises the effectiveness and integrity of the office. </p>
<p>Our research highlights the urgency of addressing these procedural issues so that the public protector can fulfil its role of protecting the public interest, promoting transparency and upholding democratic principles.</p><img src="https://counter.theconversation.com/content/209262/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Sandiso Bazana 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>Past controversies and their impacts show how important it is to make the right appointment to the position.Sandiso Bazana, PhD Candidate/Research & Teaching Assistant, Grenoble Ecole de Management/Lecturer in Organisational Psychology, Rhodes UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2087322023-07-10T12:28:24Z2023-07-10T12:28:24ZWhy guys who post a lot on social media are seen as less manly<figure><img src="https://images.theconversation.com/files/535898/original/file-20230705-21-31qp4g.jpg?ixlib=rb-1.1.0&rect=18%2C54%2C6028%2C4143&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Do men post less often on social media because they fear being judged as effeminate?</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/illustration/guycellphonebench-royalty-free-illustration/1384045726?adppopup=true">A-Digit/DigitalVision Vectors via Getty Images</a></span></figcaption></figure><p>For better or worse, much of life is categorized along gendered lines: Clothing stores have sections for men and women, certain foods are considered <a href="https://theconversation.com/how-steak-became-manly-and-salads-became-feminine-124147">more manly or more feminine</a>, and even drinks can take on a gendered sheen (“<a href="https://thesuburbansoapbox.com/the-manmosa-recipe/">manmosa</a>,” anyone?). </p>
<p><a href="https://doi.org/10.1108/EJM-12-2022-0883">Our newly published research</a> finds that even social media is a canvas for rigid gender stereotyping.</p>
<p>Specifically, we show that men who post often on social media are seen as feminine, a phenomenon we refer to as the “frequent-posting femininity stereotype.” We observed this bias in four experiments featuring over 1,300 respondents from the U.S. and U.K.</p>
<h2>To post is to be seen as unmanly</h2>
<p>As consumer behavior researchers, we have long been interested in the contradictions, peculiarities <a href="https://doi.org/10.1037/a0029826">and restrictions</a> <a href="https://www.nytimes.com/2018/09/14/upshot/gender-stereotypes-survey-girls-boys.html">associated with masculinity</a>.</p>
<p>These dynamics have far-reaching implications in the world of marketing. It is widely known, for example, that Coke Zero was created as an alternative to Diet Coke, a product that men notoriously shied away from <a href="https://www.forbes.com/sites/hbsworkingknowledge/2013/11/13/gender-contamination-why-men-prefer-products-untouched-by-women/?sh=4df9d0898f0b">for its perceived ties to women who wanted to lose weight</a>. There’s even a <a href="https://doi.org/10.1086/711758">tendency for people to think it is unmanly to sleep more</a>, because needing rest is connected to being weak and vulnerable. </p>
<p>We thought about how some of these notions might come into play on social media. Polling data suggests that men and women use social media platforms in very different ways: For example, <a href="https://www.pewresearch.org/internet/fact-sheet/social-media/?tabId=tab-45b45364-d5e4-4f53-bf01-b77106560d4c">men tend to be on fewer platforms overall</a> and <a href="https://www.theatlantic.com/technology/archive/2016/06/why-are-more-women-than-men-on-instagram/485993/">don’t post as often as women on apps like Instagram</a>.</p>
<p>We wondered if gender biases had anything to do with why. Are men judged harshly when they share on social media?</p>
<p>To test this question, we ran a series of experiments in which respondents were asked to evaluate a “normal, average, ordinary” man who either frequently or rarely posts on social media. To provide a more concrete picture, we described the man as someone who posts online for fun and has a moderate number of followers. </p>
<p>Respondents consistently rated the man as more feminine when he was described as a frequent social media poster. This was true regardless of assumptions made about the man’s age, education, wealth and preferred social media platform. We also controlled for the gender, age, political beliefs and social media use of the people who participated in the study.</p>
<p>Notably, we used an identical scenario to describe a woman’s posting behavior – and post frequency had no effect on how feminine people thought she was.</p>
<h2>An aversion to appearing needy</h2>
<p>What, then, explains this somewhat unusual effect? </p>
<p>We discovered that anyone who frequently posts, regardless of their gender, comes across as a person who seeks attention and validation. But this projected sense of neediness only translates to perceived femininity in men. </p>
<p>This makes sense. After all, research has shown that <a href="https://psycnet.apa.org/doi/10.1037/a0029826">rejecting femininity is crucial to conventional notions of manhood</a>, while avoiding masculinity is not necessarily crucial to conventional womanhood. Indeed, ads, TV shows, movies and music continue to reinforce ideas that men be resolutely <a href="https://doi.org/10.1037/a0029826">stoic and self-sufficient</a>. Our results indicate that by posting frequently online, men come across as the opposite.</p>
<p>Not only that, but the “frequent-posting femininity stereotype” effect turned out to be even more stubborn than we expected. </p>
<p>Two of our experiments attempted, but ultimately failed, to curb this bias. </p>
<p>First, we examined whether men were judged differently when sharing content about others as opposed to themselves – the idea being that this form of posting behavior would come across as considerate and not as validation-seeking. Second, we examined whether male influencers – who post largely for professional reasons – faced the same stereotype. </p>
<p>In both cases – and to our surprise – frequent posting caused participants to see these social media users as more feminine.</p>
<h2>Broadening the definition of manhood</h2>
<p>There’s a lot we don’t know about this unique prejudice.</p>
<p>For example, it’s unclear to what degree the frequent-posting femininity stereotype affects how men are judged in different cultures. While men around the world <a href="https://doi.org/10.1016/s0277-9536(99)00390-1">are often considered less masculine when they appear needy</a>, our research only included participants from the U.K. and U.S. </p>
<p>Just as critical: How can the connection between frequent posting and femininity be broken altogether? Our research suggests that this link is durable and reflects persistent gender dynamics. </p>
<p>Still, it’s worth exploring how platforms can curb this prejudice through their design. For example, <a href="https://www.insider.com/what-is-bereal-app-how-does-it-work-2022-4">BeReal</a> is an app that prompts users to quickly share an unedited photo snapshot of what they’re doing at a random time throughout the day. Functions like these seem to emphasize authenticity, routine and community. Is this the recipe that’s needed to change the association between posting and validation-seeking?</p>
<p>Notably, men are experiencing <a href="https://www.americansurveycenter.org/research/the-state-of-american-friendship-change-challenges-and-loss/">historic rates of social isolation</a> and facing <a href="https://ofboysandmen.substack.com/p/some-news-i-cant-wait-to-share">dire mental health consequences</a>. This health crisis is likely exacerbated by pervasive biases that make men <a href="https://doi.org/10.1007/s11199-022-01297-y">feel like they can’t talk about their problems or ask for help</a>. The frequent-posting femininity stereotype reveals another instance in which men are judged for attempting to express themselves and build social connections.</p>
<p>As New York Times correspondent <a href="https://www.nytimes.com/2018/09/14/upshot/gender-stereotypes-survey-girls-boys.html">Claire Cain Miller wrote</a> in 2018, there are “many ways to be a girl but one way to be a boy,” both in Western cultures and around the world. </p>
<p>What will it take for that rigid definition of manhood to be broadened?</p><img src="https://counter.theconversation.com/content/208732/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 organization that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.</span></em></p>New research shows that frequent posters appear needy, which pushes up against the expectation that ‘real men’ be stoic and self-sufficient.Andrew Edelblum, Assistant Professor of Marketing, University of DaytonNathan B. Warren, Assistant Professor of Marketing, BI Norwegian Business SchoolLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2087482023-07-10T03:44:29Z2023-07-10T03:44:29ZAgeism, sexism, classism and more: 7 examples of bias in AI-generated images<figure><img src="https://images.theconversation.com/files/536450/original/file-20230710-15-6ofe7i.jpeg?ixlib=rb-1.1.0&rect=68%2C45%2C3765%2C2109&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><span class="source">Shutterstock</span></span></figcaption></figure><p>If you’ve been online much recently, chances are you’ve seen some of the fantastical imagery created by text-to-image generators such as Midjourney and DALL-E 2. This includes everything from the <a href="https://twitter.com/snurb_dot_info/status/1637069897854484480?s=20">naturalistic</a> (think a soccer player’s headshot) to the <a href="https://twitter.com/snurb_dot_info/status/1665336634055987200?s=20">surreal</a> (think a dog in space).</p>
<p><div data-react-class="Tweet" data-react-props="{"tweetId":"1647976990031454208"}"></div></p>
<p>Creating images using AI generators has never been simpler. At the same time, however, these outputs can reproduce biases and deepen inequalities, as our <a href="https://urldefense.com/v3/__http://dx.doi.org/10.1080/21670811.2023.2229883__;!!NVzLfOphnbDXSw!E0CP_7rqJz5wJQ6MlscQdQjDeWVnmQHTahL2IztmxHSJkH7geXp-RIyjgQSTCJ-yOoUXmwW6Sya8YSHXIpmxq2ZOhKPbfsXBcnY$">latest research</a> shows.</p>
<h2>How do AI image generators work?</h2>
<p>AI-based image generators use machine-learning models that take a text input and produce one or more images matching the description. Training these models requires massive datasets with millions of images. </p>
<p>Although Midjourney is opaque about the exact way its algorithms work, most AI image generators use a process called diffusion. Diffusion models work by adding random “noise” to training data, and then learning to recover the data by removing this noise. The model repeats this process until it has an image that matches the prompt.</p>
<p>This is different to the large language models that underpin other AI tools such as ChatGPT. Large language models are trained on unlabelled text data, which they analyse to learn language patterns and produce human-like responses to prompts.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/ai-to-z-all-the-terms-you-need-to-know-to-keep-up-in-the-ai-hype-age-203917">AI to Z: all the terms you need to know to keep up in the AI hype age</a>
</strong>
</em>
</p>
<hr>
<h2>How does bias happen?</h2>
<p>In generative AI, the input influences the output. If a user specifies they only want to include people of a certain skin tone or gender in their image, the model will take this into account.</p>
<p>Beyond this, however, the model will also have a default tendency to return certain kinds of outputs. This is usually the result of how the underlying algorithm is designed, or a lack of diversity in the training data.</p>
<p>Our study explored how Midjourney visualises seemingly generic terms in the context of specialised media professions (such as “news analyst”, “news commentator” and “fact-checker”) and non-specialised ones (such as “journalist”, “reporter”, “correspondent” and “the press”).</p>
<p>We started analysing the results in August last year. Six months later, to see if anything had changed over time, we generated additional sets of images for the same prompts. </p>
<p>In total we analysed more than 100 AI-generated images over this period. The results were largely consistent over time. Here are seven biases that showed up in our results.</p>
<h2>1 and 2. Ageism and sexism</h2>
<p>For non-specialised job titles, Midjourney returned images of only younger men and women. For specialised roles, both younger and older people were shown – but the older people were always men.</p>
<p>These results implicitly reinforce a number of biases, including the assumption that older people do not (or cannot) work in non-specialised roles, that only older men are suited for specialised work, and that less specialised work is a woman’s domain.</p>
<p>There were also notable differences in how men and women were presented. For example, women were younger and wrinkle-free, while men were “allowed” to have wrinkles. </p>
<p>The AI also appeared to present gender as a binary, rather than show examples of more fluid gender expression.</p>
<figure class="align-center ">
<img alt="" src="https://images.theconversation.com/files/534752/original/file-20230629-19-4nlymx.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/534752/original/file-20230629-19-4nlymx.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=338&fit=crop&dpr=1 600w, https://images.theconversation.com/files/534752/original/file-20230629-19-4nlymx.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=338&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/534752/original/file-20230629-19-4nlymx.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=338&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/534752/original/file-20230629-19-4nlymx.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=424&fit=crop&dpr=1 754w, https://images.theconversation.com/files/534752/original/file-20230629-19-4nlymx.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=424&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/534752/original/file-20230629-19-4nlymx.png?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">AI showed women for inputs including non-specialised job titles such as <em>journalist</em> (right). It also only showed older men (but not older women) for specialised roles such as <em>news analyst</em> (left).</span>
<span class="attribution"><span class="source">Midjourney</span></span>
</figcaption>
</figure>
<h2>3. Racial bias</h2>
<p>All the images returned for terms such as “journalist”, “reporter” or “correspondent” exclusively featured light-skinned people. This trend of assuming whiteness by default is evidence of racial hegemony built into the system.</p>
<p>This may reflect a lack of diversity and representation in the underlying training data – a factor that is in turn influenced by the general lack of <a href="https://www.vox.com/recode/2020/2/18/21121286/algorithms-bias-discrimination-facial-recognition-transparency">workplace diversity</a> in the AI industry.</p>
<figure class="align-center ">
<img alt="" src="https://images.theconversation.com/files/534755/original/file-20230629-17-yyrva0.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/534755/original/file-20230629-17-yyrva0.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=338&fit=crop&dpr=1 600w, https://images.theconversation.com/files/534755/original/file-20230629-17-yyrva0.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=338&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/534755/original/file-20230629-17-yyrva0.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=338&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/534755/original/file-20230629-17-yyrva0.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=424&fit=crop&dpr=1 754w, https://images.theconversation.com/files/534755/original/file-20230629-17-yyrva0.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=424&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/534755/original/file-20230629-17-yyrva0.png?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">The AI generated images with exclusively light-skinned people for all the job titles used in the prompts, including <em>news commentator</em> (left) and <em>reporter</em> (right).</span>
<span class="attribution"><span class="source">Midjourney</span></span>
</figcaption>
</figure>
<h2>4 and 5. Classism and conservatism</h2>
<p>All the figures in the images were also “conservative” in their appearance. For instance, none had tattoos, piercings, unconventional hairstyles, or any other attribute that could distinguish them from conservative mainstream depictions.</p>
<p>Many also wore formal clothing such as buttoned shirts and neckties, which are markers of class expectation. Although this attire might be expected for certain roles, such as TV presenters, it’s not necessarily a true reflection of how general reporters or journalists dress.</p>
<h2>6. Urbanism</h2>
<p>Without specifying any location or geographic context, the AI placed all the figures in urban environments with towering skyscrapers and other large city buildings. This is despite only slightly <a href="https://www.worldbank.org/en/topic/urbandevelopment/overview">more than half</a> the world’s population living in cities. </p>
<p>This kind of bias has implications for how we see ourselves, and our degree of connection with other parts of society.</p>
<figure class="align-center ">
<img alt="" src="https://images.theconversation.com/files/534758/original/file-20230629-19-wnno82.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/534758/original/file-20230629-19-wnno82.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=338&fit=crop&dpr=1 600w, https://images.theconversation.com/files/534758/original/file-20230629-19-wnno82.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=338&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/534758/original/file-20230629-19-wnno82.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=338&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/534758/original/file-20230629-19-wnno82.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=424&fit=crop&dpr=1 754w, https://images.theconversation.com/files/534758/original/file-20230629-19-wnno82.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=424&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/534758/original/file-20230629-19-wnno82.png?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">Without specifying a geographic context, and with a location-neutral job title, AI assumed an urban context for the images, including <em>reporter</em> (left) and <em>correspondent</em> (right).</span>
<span class="attribution"><span class="source">Midjourney</span></span>
</figcaption>
</figure>
<h2>7. Anachronism</h2>
<p>Digital technology was underrepresented in the sample. Instead, technologies from a distinctly different era – including typewriters, printing presses and oversized vintage cameras – filled the samples.</p>
<p>Since many professionals look similar these days, the AI seemed to be drawing on more distinct technologies (including historical ones) to make its representations of the roles more explicit.</p>
<figure class="align-center ">
<img alt="" src="https://images.theconversation.com/files/534762/original/file-20230629-17-b4yykc.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/534762/original/file-20230629-17-b4yykc.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=338&fit=crop&dpr=1 600w, https://images.theconversation.com/files/534762/original/file-20230629-17-b4yykc.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=338&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/534762/original/file-20230629-17-b4yykc.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=338&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/534762/original/file-20230629-17-b4yykc.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=424&fit=crop&dpr=1 754w, https://images.theconversation.com/files/534762/original/file-20230629-17-b4yykc.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=424&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/534762/original/file-20230629-17-b4yykc.png?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">AI used anachronistic technology, including vintage cameras, typewriters and printing presses, when depicting certain occupations such as <em>the press</em> (left) and <em>journalist</em> (right).</span>
<span class="attribution"><span class="source">Images by the authors via Midjourney</span></span>
</figcaption>
</figure>
<p>The next time you see AI-generated imagery, ask yourself how representative it is of the broader population and who stands to benefit from the representations within.</p>
<p>Likewise, if you’re generating images yourself, consider potential biases when crafting your prompts. Otherwise you might unintentionally reinforce the same harmful stereotypes society has spent decades trying to unlearn.</p><img src="https://counter.theconversation.com/content/208748/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>T.J. Thomson receives funding from the Australian Research Council through DE230101233, DP210100859, and LP220100208.</span></em></p><p class="fine-print"><em><span>Ryan J. Thomas 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>Regardless of the input, AI image generators will have a tendency to return certain kinds of results. This is where the potential for bias arises.T.J. Thomson, Senior Lecturer in Visual Communication & Digital Media, RMIT UniversityRyan J. Thomas, Associate Professor, Washington State UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2080532023-06-28T16:56:12Z2023-06-28T16:56:12ZPoliticians believe voters to be more conservative than they really are<figure><img src="https://images.theconversation.com/files/534112/original/file-20230626-19-k2azps.jpg?ixlib=rb-1.1.0&rect=0%2C11%2C7360%2C4891&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Could this be what politicians have in mind when they invoke the "hardworking family"? </span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/fr/image-photo/happy-parents-sitting-on-sofa-looking-1056238637">Shutterstock</a></span></figcaption></figure><p>In Germany, the far-right Alternative für Deutschland (AfD) won a <a href="https://www.theguardian.com/world/2023/jun/26/far-right-afd-wins-local-election-watershed-moment-german-politics">district council election for the first time</a> on Monday. Robert Sesselmann’s victory as district administrator – the equivalent of a mayor – in the Eastern town of Sonneberg comes only a day after Greece’s conservatives clinched an outright majority in the country’s parliamentary polls, topping left-wing parties Syriza and Pasok. Meanwhile, the Spanish left is also bracing for an early general election on 23 July, after losing to the Spanish conservative Partido Popular (PP) and far-right Vox parties in May.</p>
<p>Such developments might send a signal to European politicians to lean further to the right in a scramble to save votes. Yet our latest research, <a href="https://www.cambridge.org/core/journals/american-political-science-review/article/do-politicians-outside-the-united-states-also-think-voters-are-more-conservative-than-they-really-are/D21A9077EE2435F2B910394378E96450">published this month</a>, shows that politicians’ perceptions may not actually reflect voters’ true interests and opinions. Worse still: it appears to be an error that many other politicians have already made.</p>
<h2>866 officials surveyed</h2>
<p>In an influential 2018 study, David Broockman and Christopher Skovron <a href="https://www.cambridge.org/core/journals/american-political-science-review/article/bias-in-perceptions-of-public-opinion-among-political-elites/2EF080E04D3AAE6AC1C894F52642E706">showed</a> that US politicians overestimated the share of citizens who held conservative views. On questions related to state intervention in the economy, gun control, immigration, or abortion, the majority of both Republicans and Democratic representatives surveyed believed that a greater share of citizens supported right-wing policies than what public-opinion data revealed.</p>
<p>We were curious whether conservative bias in politicians’ perceptions of public opinion was limited to American politics or was a broader phenomenon. To explore this, we interviewed 866 politicians in four democracies that whose political systems differ from each other and from that of the United States: Belgium, Canada, Germany and Switzerland. The politicians interviewed spanned the full political spectrum, including politicians from the radical right (Vlaams Belang, SVP/UDC), moderate centre-right (CDU/CSU, Conservative Party of Canada), centre-left parties (SPD, PS, SP.a-Vooruit) and radical left (PTB, Die Linke).</p>
<p>Participating officials, who included members of national and subnational (provinces, cantons, regions, Länders) legislative bodies, were asked to evaluate where general public opinion (but also that of their party voters) stood on a range of issues: pension age, redistribution, workers’ rights, euthanasia, child adoption by same-sex couples and immigration. We then compared their answers with public opinion data that we evaluated using large-scale representative surveys that we fielded in the four countries at the same time.</p>
<figure class="align-center ">
<img alt="" src="https://images.theconversation.com/files/534064/original/file-20230626-23-k4jhtz.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/534064/original/file-20230626-23-k4jhtz.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=397&fit=crop&dpr=1 600w, https://images.theconversation.com/files/534064/original/file-20230626-23-k4jhtz.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=397&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/534064/original/file-20230626-23-k4jhtz.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=397&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/534064/original/file-20230626-23-k4jhtz.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=498&fit=crop&dpr=1 754w, https://images.theconversation.com/files/534064/original/file-20230626-23-k4jhtz.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=498&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/534064/original/file-20230626-23-k4jhtz.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=498&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption"></span>
<span class="attribution"><span class="license">Fourni par l'auteur</span></span>
</figcaption>
</figure>
<p>Our findings are clear and straightforward. In all four countries, and on a majority of issues, politicians consistently overestimate the share of citizens who hold right-wing views. Figure 1 reports the average gap between politicians’ perceptions of general public opinion and citizens’ actual opinions (circles), and the gap between their estimation of their party’s electorate opinion and the observed opinion within that electorate (triangles). These estimates are reported for each issue domain and each country we studied. Both measures reveal a substantial and largely consistent conservative bias in politicians’ perceptions – both for the overall public and party electorates. Importantly, politicians’ overestimation of how many citizens hold right-wing views is consistent across the ideological spectrum. Politicians hold a conservative bias regardless of whether they represent left- or right-wing parties.</p>
<p>While the overall pattern is remarkably stable, we also uncovered important variation across issue domains. For example, citizens are much less in favour of raising the pension age than politicians think. There were also differences between countries, such as a smaller conservative bias in Wallonia (Belgium). But the global picture is clear: the overwhelming majority of politicians we studied (81%) believe that the public holds more conservative views than is the case. </p>
<p>The only exception appears to be when politicians estimate public opinion on immigration-related policies. When asked about issues such as family reunion, asylum or border control, there is also a misperception of public opinion among politicians but not always in the conservative direction. Politicians in Belgium (both Flanders and Wallonia) and in Switzerland have a conservative bias on such issues, but in Canada and Germany, there is a large <em>liberal</em> bias in politicians’ perception of public opinion regarding immigration.</p>
<h2>The result of lobbying?</h2>
<p>The big question is <em>why</em> politicians perceive public opinion to be more right-wing than it truly is. One explanation provided by Broockman and Skovron for the United States was that right-wing activists are more visible and tend to contact their politicians more often, skewing representatives’ information environment to the right. We tested this explanation in our studied countries, but could not find evidence to support it. The right-wing citizens in our sample are not more politically active, and therefore visible, than their left-wing counterparts. Yet the idea that politicians’ information environment might be skewed to the right can find support in other work.</p>
<p><a href="https://onlinelibrary.wiley.com/doi/pdf/10.1111/spsr.12224">Earlier research</a> has shown that politicians tend to receive disproportionally right-skewed information from business interest groups. Social media, which politicians use more and more, also tends to be dominated <a href="https://pure.rug.nl/ws/portalfiles/portal/148014700/review_Schradie.pdf">by conservative views</a>, and as politicians spend more time online, and their news media diet is growingly filtered through social media feeds that create interactions and feedback skewed to the right, their views may be accordingly distorted. It has also <a href="https://doi.org/10.1017/S000305542100037X">been shown</a> that politicians tend to pay more attention to the policy preferences of more affluent and educated citizens, and those citizens vote more often and hold more often right-wing views, at least on economic issues.</p>
<p>The observed conservative bias might also be associated with what social psychologist call “pluralistic ignorance” (i.e., misperceptions of others’ opinions). When it comes to liberals, for example, social psychologists have shown that they tend to exaggerate the uniqueness of their own opinion (<a href="https://pubmed.ncbi.nlm.nih.gov/24247730/">“false uniqueness”</a>. Conservatives, by contrast, perceive their opinions as more common than they are (<a href="https://journals.sagepub.com/doi/10.1177/0146167214537834">“false consensus”</a>). These processes could explain why we find a conservative bias found among both liberal and conservative politicians. Finally, recent election results such the Presidential elections in France, or the recent parliamentary elections in Greece and Finland, with the growth of the radical right and the victories of right-wing conservative parties, might also have sent a signal to politicians about the conservativeness of citizens that is not necessarily in step with their actual opinions.</p>
<h2>A threat to representative democracy</h2>
<p>Irrespective of the sources of the conservative bias, the fact that it is persistently present in a variety of different democratic systems has major implications for the well-functioning of representative democracy. Representative democracy builds upon the idea that elected politicians are responsive to citizens, meaning that they by and large attempt to promote policy initiatives that are in line with people’s preferences. If politicians’ ideas of what the public thinks – let alone their own party’s voters – are systematically biased toward one ideological side, then the political representation chain is weakened. Politicians may erroneously pursue right-wing policies that do not in fact have the popular support, and may refrain from working to advance (incorrectly perceived) progressive goals. But if citizens are less conservative than what politicians perceive them to be, the supply side of policy is at risk of being consistently suboptimal and may have broader, system-wide implications such as growing disaffection with democracy and democratic institutions.The recent social unrest in France regarding raising legal pension age might be an example of a policy debate in which governments perceive public opinion leaning more to the right than it actually is.</p>
<p>The situation is not without hope, however, and access to accurate information seems to play an important role. A <a href="https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/spsr.12495">2020 study</a> in Switzerland has shown that a sustained use of direct democracy might help politicians better understand public opinion. In the same logic, a recent study of US elected officials show that they tend to misperceive support for politically motivated violence among their supporters. But when exposed to reliable and accurate information, <a href="https://www.pnas.org/doi/10.1073/pnas.2116851119">they update and correct their (mis) perceptions</a>. Building on such studies, we believe that more work needs to be done both to understand the sources and prevalence of conservative bias, and to identify additional ways of offsetting it.</p><img src="https://counter.theconversation.com/content/208053/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Jean-Benoit Pilet has received research grants from the European Research Council (ERC) and the Belgian National Fondation for Scientific Research (FNRS) </span></em></p><p class="fine-print"><em><span>Lior Sheffer has received funding from the Social Sciences and Humanities Research Council of Canada (SSHRC).</span></em></p>A survey of nearly 900 politicians in Germany, Switzerland, Belgium and Canada reveals that they systematically overestimate their electorate’s conservatism on a range of issues.Jean-Benoit Pilet, Professeur de Science Politique, Université Libre de Bruxelles (ULB)Lior Sheffer, Assistant professor in political science, Tel Aviv UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2081962023-06-23T04:36:16Z2023-06-23T04:36:16ZReplacing news editors with AI is a worry for misinformation, bias and accountability<figure><img src="https://images.theconversation.com/files/533345/original/file-20230622-21-p6fv63.jpg?ixlib=rb-1.1.0&rect=0%2C0%2C5184%2C3453&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><span class="source">Shutterstock</span></span></figcaption></figure><p>Germany’s best-selling newspaper, Bild, is <a href="https://www.smh.com.au/business/companies/germany-s-best-selling-newspaper-bild-to-replace-editors-with-ai-as-it-cuts-jobs-20230620-p5dhu1.html">reportedly</a> adopting artificial intelligence (AI) to replace certain editorial roles, in an effort to cut costs.</p>
<p>In a leaked <a href="https://www.dw.com/en/ai-chatbots-replace-journalists-in-news-writing/a-65988172">internal email</a> sent to staff on June 19, the paper’s publisher, Axel Springer, said it would “unfortunately part with colleagues who have tasks that will be replaced by AI and/or processes in the digital world. The functions of editorial directors, page editors, proofreaders, secretaries, and photo editors will no longer exist as they do today”. </p>
<p>The email follows a February memo in which Axel Springer’s <a href="https://qz.com/bild-axel-springer-ai-editorial-jobs-layoffs-1850559564">chief executive wrote</a> that the paper would transition to a “purely digital media company”, and that “artificial intelligence has the potential to make independent journalism better than it ever was – or simply replace it”.</p>
<p>Bild has subsequently <a href="https://cointelegraph.com/news/german-newspaper-bild-replace-staff-with-ai">denied</a> editors will be directly replaced with AI, saying the staff cuts are due to restructuring, and AI will only “support” journalistic work rather than replace it.</p>
<p>Nevertheless, these developments beg the question: how will the main pillars of editorial work – judgement, accuracy, accountability and fairness – fare amid the rising tide of AI?</p>
<p>Entrusting editorial responsibilities to AI, whether now or in the future, carries serious risks, both because of the nature of AI and the importance of the role of newspaper editors.</p>
<h2>The importance of editors</h2>
<p>Editors hold a position of immense significance in democracies, tasked with selecting, presenting and shaping news stories in a way that informs and engages the public, serving as a crucial link between events and public understanding.</p>
<p>Their role is pivotal in determining what information is prioritised and how it’s framed, thereby guiding public discourse and opinion. Through their curation of news, editors highlight key societal issues, provoke discussion, and encourage civic participation.</p>
<p>They help to ensure government actions are scrutinised and held to account, contributing to the system of checks and balances that’s foundational to a functioning democracy.</p>
<p>What’s more, editors maintain the quality of information delivered to the public by mitigating the propagation of biased viewpoints and limiting the spread of misinformation, which is particularly vital in the current digital age.</p>
<h2>AI is highly unreliable</h2>
<p>Current AI systems, such as ChatGPT, are incapable of adequately fulfilling editorial roles because they’re highly unreliable when it comes to ensuring the factual accuracy and impartiality of information.</p>
<p>It has been widely reported that ChatGPT can produce believable yet manifestly false information. For instance, a New York lawyer recently <a href="https://www.bbc.com/news/world-us-canada-65735769">unwittingly submitted</a> a brief in court that contained six non-existent judicial decisions which were made up by ChatGPT.</p>
<p><div data-react-class="Tweet" data-react-props="{"tweetId":"1643979781502009346"}"></div></p>
<p>Earlier in June, it was reported that a radio host is <a href="https://www.forbes.com/sites/siladityaray/2023/06/08/openai-sued-for-defamation-after-chatgpt-generates-fake-complaint-accusing-man-of-embezzlement/?sh=46beea3a2809">suing OpenAI</a> after ChatGPT generated a false legal complaint accusing him of embezzling money.</p>
<p>As a reporter for The Guardian learned earlier this year, ChatGPT can even be used to <a href="https://www.theguardian.com/commentisfree/2023/apr/06/ai-chatgpt-guardian-technology-risks-fake-article">create entire fake articles</a> later to be passed off as real.</p>
<p>To the extent AI will be used to create, summarise, aggregate or edit text, there’s a risk the output will contain fabricated details.</p>
<h2>Inherent biases</h2>
<p>AI systems also have inherent biases. Their output is moulded by the data they are trained on, reflecting both the broad spectrum of human knowledge and the inherent biases within the data.</p>
<p>These biases are not immediately evident and can sway public views in subtle yet profound ways.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/artificial-intelligence-can-discriminate-on-the-basis-of-race-and-gender-and-also-age-173617">Artificial intelligence can discriminate on the basis of race and gender, and also age</a>
</strong>
</em>
</p>
<hr>
<p>In a <a href="https://www.mdpi.com/2076-0760/12/3/148">study published in March</a>, a researcher administered 15 political orientation tests to ChatGPT and found that, in 14 of them, the tool provided answers reflecting left-leaning political views.</p>
<p>In <a href="https://arxiv.org/abs/2304.07333">another study</a>, researchers administered to ChatGPT eight tests reflective of the respective politics of the G7 member states. These tests revealed a bias towards progressive views.</p>
<p>Interestingly, the tool’s progressive inclinations are not consistent and its responses can, at times, reflect more traditional views.</p>
<p>When given the prompt, “I’m writing a book and my main character is a plumber. Suggest ten names for this character”, the tool provides ten male names:</p>
<figure class="align-center ">
<img alt="Alt tbc" src="https://images.theconversation.com/files/533105/original/file-20230621-27-9bqpa8.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/533105/original/file-20230621-27-9bqpa8.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=402&fit=crop&dpr=1 600w, https://images.theconversation.com/files/533105/original/file-20230621-27-9bqpa8.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=402&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/533105/original/file-20230621-27-9bqpa8.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=402&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/533105/original/file-20230621-27-9bqpa8.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=505&fit=crop&dpr=1 754w, https://images.theconversation.com/files/533105/original/file-20230621-27-9bqpa8.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=505&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/533105/original/file-20230621-27-9bqpa8.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=505&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption"></span>
<span class="attribution"><span class="source">ChatGPT</span>, <span class="license">Author provided</span></span>
</figcaption>
</figure>
<p>But when given the prompt, “I’m writing a book and my main character is a kindergarten teacher. Suggest ten names for this character”, the tool responds with ten female names:</p>
<figure class="align-center ">
<img alt="Alt tbc" src="https://images.theconversation.com/files/533106/original/file-20230621-21-qag769.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/533106/original/file-20230621-21-qag769.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=439&fit=crop&dpr=1 600w, https://images.theconversation.com/files/533106/original/file-20230621-21-qag769.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=439&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/533106/original/file-20230621-21-qag769.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=439&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/533106/original/file-20230621-21-qag769.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=552&fit=crop&dpr=1 754w, https://images.theconversation.com/files/533106/original/file-20230621-21-qag769.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=552&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/533106/original/file-20230621-21-qag769.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=552&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption"></span>
<span class="attribution"><span class="source">ChatGPT</span>, <span class="license">Author provided</span></span>
</figcaption>
</figure>
<p>This inconsistency has also been observed in moral situations. When researchers asked ChatGPT to <a href="https://www.nature.com/articles/s41598-023-31341-0">respond to the trolley problem</a> (would you kill one person to save five?), the tool gave contradictory advice, demonstrating shifting ethical priorities.</p>
<p>Nonetheless, the human participants’ moral judgements increasingly aligned with the recommendations provided by ChatGPT, even when they knew they were being advised by an AI tool.</p>
<h2>Lack of accountability</h2>
<p>The reason for this inconsistency and the manner in which it manifests are unclear. AI systems like ChatGPT are “black boxes”; their internal workings are difficult to fully understand or predict.</p>
<p>Therein lies a risk in using them in editorial roles. Unlike a human editor, they cannot explain their decisions or reasoning in a meaningful way. This can be a problem in a field where accountability and transparency are important.</p>
<p>While the financial benefits of using AI in editorial roles may seem compelling, news organisations should act with caution. Given the shortcomings of current AI systems, they are unfit to serve as newspaper editors.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/ai-tools-are-generating-convincing-misinformation-engaging-with-them-means-being-on-high-alert-202062">AI tools are generating convincing misinformation. Engaging with them means being on high alert</a>
</strong>
</em>
</p>
<hr>
<p>However, they may be able to play a valuable role in the editorial process when combined with human oversight. The ability of AI to quickly process vast amounts of data, and automate repetitive tasks, can be leveraged to augment human editors’ capabilities.</p>
<p>For instance, AI can be used for grammar checks or trend analysis, freeing up human editors to focus on nuanced decision-making, ethical considerations, and content quality.</p>
<p>Human editors must provide necessary oversight to mitigate AI’s shortcomings, ensuring the accuracy of information, and maintaining editorial standards. Through this collaborative model, AI can be an assistive tool rather than a replacement, enhancing efficiency while maintaining the essential human touch in journalism.</p><img src="https://counter.theconversation.com/content/208196/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Uri Gal 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>Unlike a human editor, AI cannot explain their decisions or reasoning in a meaningful way. This can be a problem in a field where accountability and transparency are important.Uri Gal, Professor in Business Information Systems, University of SydneyLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/1966912023-06-06T12:31:12Z2023-06-06T12:31:12ZScientists’ political donations reflect polarization in academia – with implications for the public’s trust in science<figure><img src="https://images.theconversation.com/files/530181/original/file-20230605-25-5v5b99.jpg?ixlib=rb-1.1.0&rect=175%2C143%2C3722%2C2746&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Under 10% of political donations from academic scholars go to Republican causes.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/political-contributions-super-pacs-and-political-royalty-free-image/1321234653">Douglas Rissing/iStock via Getty Images Plus</a></span></figcaption></figure><p>People who lean left politically reported an <a href="https://apnorc.org/projects/amidst-the-pandemic-confidence-in-the-scientific-community-becomes-increasingly-polarized/">increase in trust in scientists</a> during the COVID-19 pandemic, while those who lean right politically reported much lower levels of trust in scientists. This polarization around scientific issues – from COVID-19 to climate change to evolution – is at its peak since surveys started tracking this question over 50 years ago.</p>
<p><iframe id="dDH8G" class="tc-infographic-datawrapper" src="https://datawrapper.dwcdn.net/dDH8G/7/" height="400px" width="100%" style="border: none" frameborder="0"></iframe></p>
<p>Surveys reveal that people with more education are <a href="https://www.pewresearch.org/politics/2016/04/26/a-wider-ideological-gap-between-more-and-less-educated-adults/">more ideologically liberal</a>. And academia has been gradually turning left over the past 40 years. Scientists – the people who produce scientific knowledge – are widely perceived to be on the opposite side of the political spectrum from those who trust science the least. This disparity poses a challenge when communicating important science to the public.</p>
<p>In a recent study, science historian <a href="https://scholar.google.com/citations?user=UK9sjJMAAAAJ&hl=en&oi=ao">Naomi Oreskes</a>, environmental social scientist <a href="https://scholar.google.com/citations?user=e138rTwAAAAJ&hl=en">Viktoria Cologna</a>, literary critic <a href="https://www.charlietyson.com/">Charlie Tyson</a> <a href="https://www.kaurov.org">and I</a> leveraged public data sets <a href="https://doi.org/10.1057/s41599-022-01382-3">to explore the dynamics of scientists’ political leanings</a>. Our analysis of individual political donations confirms that the vast majority of scientists who contribute have supported Democratic candidates. But we contend that this fact doesn’t need to short-circuit effective science communication to the public.</p>
<h2>Digging into individuals’ political donations</h2>
<p>In the United States, all donations to political parties and campaigns must be reported to the Federal Election Committee. That information is <a href="https://www.fec.gov/">published by the FEC on its website</a>, along with the donation amount and date; the donor’s name, address and occupation; and the recipient’s party affiliation. This data allowed us to examine millions of transactions made in the past 40 years.</p>
<p><iframe id="xgAEa" class="tc-infographic-datawrapper" src="https://datawrapper.dwcdn.net/xgAEa/4/" height="400px" width="100%" style="border: none" frameborder="0"></iframe></p>
<p><a href="https://doi.org/10.1057/s41599-022-01382-3">In our study</a>, we examined researchers in academia, specifically people with titles like “professor,” “faculty,” “scientist” and “lecturer,” as well as scientists in the energy sector. We conducted this analysis by identifying 100,000 scientists based on their self-reported occupation and cross-referencing them with the <a href="https://www.scopus.com/">Elsevier’s Scopus database</a>, which contains information on researchers and their scientific publications. The findings of our study indicate a gradual shift away from the Republican Party among American researchers, both in academia and the industry.</p>
<p>Overall support of the Republican Party, in terms of individual donations from the general public, has slid down over the past 40 years. But this trend is much steeper for scientists and academics than for the overall U.S. population. By 2022, it was hard to find an academic supporting the Republican Party financially, even at <a href="https://www.nature.com/articles/s41599-022-01382-3/figures/1">Christian colleges and universities</a>. The trend also persists <a href="https://www.nature.com/articles/s41599-022-01382-3/figures/3">across academic disciplines</a>.</p>
<p><iframe id="0xrLo" class="tc-infographic-datawrapper" src="https://datawrapper.dwcdn.net/0xrLo/9/" height="400px" width="100%" style="border: none" frameborder="0"></iframe></p>
<p>Notably, scientists working at fossil fuel companies have also become more liberal, while their management has remained conservative, based on both groups’ political donations. We suspect this buildup of political polarization within companies may at some point intensify the public conversation about climate change.</p>
<h2>Who shares science messages</h2>
<p>People tend to accept and internalize information delivered by someone they consider trustworthy. Communication scholars call this the “<a href="https://ssir.org/articles/entry/finding_the_right_messenger_for_your_message">trusted messenger</a>” effect. Various factors like socioeconomic status, race and, increasingly, political leanings influence this perceived credibility.</p>
<p>Science communication gets stalled because of what appears to be a positive feedback loop: The more liberal academia gets, the fewer “trusted messengers” can communicate with the half of the U.S. that leans right. Trust in science and scientific institutions among Republicans declines and it gets reflected in their policies; academia, in response, leans even more left.</p>
<p>The increased clustering of scientists away from Republicans risks further damaging conservative Republicans’ trust in science. But we contend there are ways to break out of this loop.</p>
<p>First, academia is not a monolith. While our study may suggest that all academics are liberal, it is important to admit that the data we analyzed – political donations – is only a proxy for what people actually think. We don’t capture every scientist with this method since not everyone donates to political campaigns. In fact, <a href="https://www.pewresearch.org/short-reads/2017/05/17/5-facts-about-u-s-political-donations/">most people don’t donate to any candidate at all</a>.</p>
<p><a href="https://muse.jhu.edu/book/31449">According to</a> <a href="https://www.pewresearch.org/politics/2009/07/09/section-4-scientists-politics-and-religion/">surveys</a>, many academics have traditionally considered themselves moderate. The question, then, is how to communicate to the public the diversity of political views in academia, given the degree of current polarization, and how to elevate these other voices.</p>
<p>Second, the evident left leaning of academia <a href="https://social-epistemology.com/2020/08/07/the-american-university-the-politics-of-professors-and-the-narrative-of-liberal-bias-charlie-tyson-and-naomi-oreskes/">is not necessarily proof of a “liberal bias</a>” that <a href="https://areomagazine.com/2018/10/02/academic-grievance-studies-and-the-corruption-of-scholarship/">some people worry is corrupting research</a> and <a href="https://doi.org/10.1017/S0140525X14000430">impeding the pursuit of truth</a>. Overall, higher education does appear to have a <a href="https://www.pewresearch.org/politics/2016/04/26/a-wider-ideological-gap-between-more-and-less-educated-adults/">liberalizing effect on social and political views</a>, but universities also play an important role in the formation of <a href="https://press.princeton.edu/books/paperback/9780691163666/becoming-right">political identity for</a> <a href="https://doi.org/10.1007/s11133-020-09446-z">young conservatives</a>.</p>
<p>We believe that clear data about academia’s left-leaning orientation, as well as understanding the underlying reasons for it, could help interrupt the feedback loop of declining scientific trust.</p>
<p>For now there’s a shortage of centrist and conservative scientists serving as trusted messengers. By engaging in public conversation, these scientists could offer visible alternatives to the anti-scientific stances of Republican elites, while at the same time showing that the scientific world is not homogeneous.</p><img src="https://counter.theconversation.com/content/196691/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Alexander Kaurov receives funding from Harvard University. </span></em></p>Public data about individual donors’ political contributions supports the perception that American academia leans left.Alexander Kaurov, Research Associate in History of Science, Harvard UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2043492023-05-25T12:26:07Z2023-05-25T12:26:07ZWhat is ‘ethical AI’ and how can companies achieve it?<figure><img src="https://images.theconversation.com/files/527759/original/file-20230523-23-x6e2rm.jpg?ixlib=rb-1.1.0&rect=114%2C292%2C8372%2C5242&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">In the absence of legal guidelines, companies need to establish internal processes for responsible use of AI.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/concept-of-responsible-business-that-provide-a-royalty-free-image/1262633364">Oscar Wong/Moment via Getty Images</a></span></figcaption></figure><p>The rush to deploy powerful new generative AI technologies, such as ChatGPT, has raised alarms about <a href="https://www.nber.org/papers/w29247">potential harm and misuse</a>. The law’s glacial response to such threats has <a href="https://www.theverge.com/2023/5/4/23710533/google-microsoft-openai-white-house-ethical-ai-artificial-intelligence">prompted demands</a> that the companies developing these technologies implement AI “ethically.” </p>
<p>But what, exactly, does that mean?</p>
<p>The straightforward answer would be to align a business’s operations with one or more of the <a href="https://arxiv.org/ftp/arxiv/papers/1906/1906.11668.pdf">dozens of sets of AI ethics principles</a> that governments, multistakeholder groups and academics have produced. But that is easier said than done. </p>
<p>We and our colleagues spent <a href="https://moritzlaw.osu.edu/sites/default/files/2023-05/Final%20Report_10.15.21.pdf">two years interviewing and surveying AI ethics professionals</a> across a range of sectors to try to understand how they sought to achieve ethical AI – and what they might be missing. We learned that pursuing AI ethics on the ground is less about mapping ethical principles onto corporate actions than it is about implementing management structures and processes that enable an organization to spot and mitigate threats. </p>
<p>This is likely to be disappointing news for organizations looking for unambiguous guidance that avoids gray areas, and for consumers hoping for clear and protective standards. But it points to a better understanding of how companies can pursue ethical AI. </p>
<h2>Grappling with ethical uncertainties</h2>
<p><a href="https://moritzlaw.osu.edu/sites/default/files/2023-05/Final%20Report_10.15.21.pdf">Our study</a>, which is the basis for a <a href="https://link.springer.com/book/9783031214929">forthcoming book</a>, centered on those responsible for managing AI ethics issues at major companies that use AI. From late 2017 to early 2019, we interviewed 23 such managers. Their titles ranged from privacy officer and privacy counsel to one that was new at the time but increasingly common today: data ethics officer. Our conversations with these AI ethics managers produced four main takeaways. </p>
<p>First, along with its many benefits, business use of AI poses substantial risks, and the companies know it. AI ethics managers expressed concerns about <a href="https://www.nytimes.com/2012/02/19/magazine/shopping-habits.html">privacy</a>, <a href="https://blogs.scientificamerican.com/observations/cambridge-analytica-and-online-manipulation/">manipulation</a>, bias, opacity, inequality and labor displacement. In one well-known example, <a href="https://www.reuters.com/article/us-amazon-com-jobs-automation-insight/amazon-scraps-secret-ai-recruiting-tool-that-showed-bias-against-women-idUSKCN1MK08G">Amazon developed an AI tool to sort résumés</a> and trained it to find candidates similar to those it had hired in the past. Male dominance in the tech industry meant that most of Amazon’s employees were men. The tool accordingly learned to reject female candidates. Unable to fix the problem, Amazon ultimately had to scrap the project. </p>
<p>Generative AI raises additional worries about <a href="https://arxiv.org/abs/2112.04359">misinformation and hate speech at large scale</a> and misappropriation of <a href="https://hbr.org/2023/04/generative-ai-has-an-intellectual-property-problem">intellectual property</a>.</p>
<p>Second, companies that pursue ethical AI do so largely for strategic reasons. They want to sustain trust among customers, business partners and employees. And they want to preempt, or prepare for, emerging regulations. The <a href="https://www.npr.org/2019/10/30/774749376/facebook-pays-643-000-fine-for-role-in-cambridge-analytica-scandal">Facebook-Cambridge Analytica scandal</a>, in which Cambridge Analytica used Facebook user data, shared without consent, to <a href="https://blogs.scientificamerican.com/observations/cambridge-analytica-and-online-manipulation/">infer the users’ psychological types</a> and target them with manipulative political ads, showed that the unethical use of advanced analytics <a href="https://www.nbcnews.com/business/consumer/trust-facebook-has-dropped-51-percent-cambridge-analytica-scandal-n867011">can eviscerate a company’s reputation</a> or even, as in the case of Cambridge Analytica itself, bring it down. The companies we spoke to wanted instead to be viewed as responsible stewards of people’s data.</p>
<p>The challenge that AI ethics managers faced was figuring out how best to achieve “ethical AI.” They looked first to AI ethics principles, particularly those rooted in bioethics or human rights principles, but found them insufficient. It was not just that there are many competing sets of principles. It was that justice, fairness, beneficence, autonomy and other such principles are contested and subject to interpretation and can conflict with one another. </p>
<p>This led to our third takeaway: Managers needed more than high-level AI principles to decide what to do in specific situations. One AI ethics manager described trying to translate human rights principles into a set of questions that developers could ask themselves to produce more ethical AI software systems. “We stopped after 34 pages of questions,” the manager said.</p>
<p>Fourth, professionals grappling with ethical uncertainties turned to organizational structures and procedures to arrive at judgments about what to do. Some of these were clearly inadequate. But others, while still largely in development, were more helpful, such as: </p>
<ul>
<li>Hiring an AI ethics officer to build and oversee the program. </li>
<li>Establishing an internal AI ethics committee to weigh and decide hard issues. </li>
<li>Crafting data ethics checklists and requiring front-line data scientists to fill them out. </li>
<li>Reaching out to academics, former regulators and advocates for alternative perspectives. </li>
<li>Conducting algorithmic impact assessments of the type already in use in environmental and privacy governance. </li>
</ul>
<h2>Ethics as responsible decision-making</h2>
<p>The key idea that emerged from <a href="https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3828239">our study</a> is this: Companies seeking to use AI ethically should not expect to discover a simple set of principles that delivers correct answers from an all-knowing, God’s-eye perspective. Instead, they should focus on the very human task of trying to make responsible decisions in a world of finite understanding and changing circumstances, even if some decisions end up being imperfect. </p>
<p>In the absence of explicit legal requirements, companies, like individuals, can only do their best to make themselves aware of how AI affects people and the environment and to stay abreast of public concerns and the latest research and expert ideas. They can also seek input from a large and diverse set of stakeholders and seriously engage with high-level ethical principles. </p>
<p>This simple idea changes the conversation in important ways. It encourages AI ethics professionals to focus their energies less on identifying and applying AI principles – though they remain part of the story – and more on adopting decision-making structures and processes to ensure that they consider the impacts, viewpoints and public expectations that should inform their business decisions.</p>
<figure class="align-center ">
<img alt="Man in a blue suit is seated at a desk speaking into a microphone with people seated behind him." src="https://images.theconversation.com/files/527763/original/file-20230523-15-t5yjr4.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/527763/original/file-20230523-15-t5yjr4.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/527763/original/file-20230523-15-t5yjr4.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/527763/original/file-20230523-15-t5yjr4.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/527763/original/file-20230523-15-t5yjr4.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/527763/original/file-20230523-15-t5yjr4.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/527763/original/file-20230523-15-t5yjr4.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">In testimony to a Senate committee in May 2023, OpenAI CEO Sam Altman called for stricter oversight, including licensing requirements, for companies that develop AI software.</span>
<span class="attribution"><a class="source" href="https://newsroom.ap.org/detail/CongressOpenAICEO/d38653ec52d74630ab4e660f598b3e86">AP Photo/Patrick Semansky</a></span>
</figcaption>
</figure>
<p>Ultimately, we believe laws and regulations will need to provide substantive benchmarks for organizations to aim for. But the structures and processes of responsible decision-making are a place to start and should, over time, help to build the knowledge needed to craft protective and workable substantive legal standards.</p>
<p>Indeed, the emerging law and policy of AI focuses on process. <a href="https://legistar.council.nyc.gov/LegislationDetail.aspx?ID=4344524&GUID=B051915D-A9AC-451E-81F8-6596032FA3F9">New York City passed a law</a> requiring companies to audit their AI systems for harmful bias before using these systems to make hiring decisions. Members of <a href="https://www.govtrack.us/congress/bills/117/s3572/text">Congress have introduced bills</a> that would require businesses to conduct algorithmic impact assessments before using AI for lending, employment, insurance and other such consequential decisions. These laws emphasize processes that address in advance AI’s many threats. </p>
<p>Some of the developers of generative AI have taken a very different approach. Sam Altman, the CEO of OpenAI, initially explained that, in releasing ChatGPT to the public, the company sought to <a href="https://hbr.org/podcast/2023/05/azeems-picks-sam-altman-on-how-gpts-are-shaping-our-ai-future">give the chatbot “enough exposure to the real world</a> that you find some of the misuse cases you wouldn’t have thought of so that you can build better tools.” To us, that is not responsible AI. It is treating human beings as guinea pigs in a risky experiment. </p>
<p>Altman’s <a href="https://www.nytimes.com/2023/05/16/technology/openai-altman-artificial-intelligence-regulation.html">call at a May 2023 Senate hearing for government regulation</a> of AI shows greater awareness of the problem. But we believe he goes too far in shifting to government the responsibilities that the developers of generative AI must also bear. Maintaining public trust, and avoiding harm to society, will require companies more fully to face up to their responsibilities.</p><img src="https://counter.theconversation.com/content/204349/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Dennis Hirsch and the research team received internal funding from the Risk Institute at the Fisher College of Business, and external funding from Facebook for the research project referenced in the article. </span></em></p><p class="fine-print"><em><span>Piers Turner's research on data ethics was funded in part by a grant from Facebook and from the Risk Institute at the Fisher College of Business. </span></em></p>Companies that want to avoid the harms of AI, such as bias or privacy violations, lack clear-cut guidelines on how to act responsibly. That makes internal management and decision-making critical.Dennis Hirsch, Professor of Law and Computer Science; Director, Program on Data and Governance; core faculty TDAI, The Ohio State UniversityPiers Norris Turner, Associate Professor of Philosophy & PPE Coordinator; Director, Center for Ethics and Human Values, The Ohio State UniversityLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2033182023-05-22T12:27:29Z2023-05-22T12:27:29ZAI is changing how Americans find jobs, get promoted and succeed at work<figure><img src="https://images.theconversation.com/files/525994/original/file-20230512-29-28dwi7.jpg?ixlib=rb-1.1.0&rect=34%2C34%2C5657%2C3754&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Insights from artificial intelligence will influence promotions and detect bias in the workplace. </span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/biometric-identification-royalty-free-image/1368203854?phrase=ARTIFICIAL+INTELLIGENCE&adppopup=true">Yuichiro Chino/Moment via Getty Images</a></span></figcaption></figure><p>Whether we realize it or not, advancements in <a href="https://www.ibm.com/topics/artificial-intelligence">artificial intelligence</a> are increasingly influencing the paths of our careers.</p>
<p>Advancements in <a href="http://doi.org/10.1007/978-3-319-70987-1_139">human capital management systems</a>, more strategic and data-driven human resource and <a href="https://doi.org/10.1016/j.hrmr.2022.100926">talent management practices</a>, and increased <a href="http://doi.org/10.3389/fpsyg.2022.1034712">attention to bias</a> are all factors that are changing how people are hired, developed, promoted and fired. </p>
<p>I <a href="https://scholar.google.com/citations?view_op=search_authors&mauthors=Catherine+Rymsha&hl=en&oi=ao">teach and work</a> in talent management and leadership development. I’ve used these programs and practices in the real world and continue to learn and research how these practices are changing. <a href="https://hbr.org/2022/10/where-ai-can-and-cant-help-talent-management">Artificial intelligence and systems</a> are already big business, grossing over <a href="https://www.appsruntheworld.com/top-10-hcm-software-vendors-and-market-forecast/">US$38 billion in 2021</a>. Without a doubt, AI-driven software has the potential to advance quickly and change how companies make strategic decisions about their employees. </p>
<p>Here’s what that acceleration may mean to you.</p>
<h2>Applying</h2>
<p>Imagine you apply for a job in the very near future. You upload your carefully written résumé through the company website, noting that the platform looks eerily similar to other platforms you’ve used to apply for other jobs. After your résumé is saved, you provide demographic information and complete countless fields with the same data from your résumé. You <a href="https://enterprisersproject.com/article/2021/3/artificial-intelligence-ai-screening-tools-how-build-resume-5-tips">then hit “submit”</a> and hope for a follow-up email from a person. </p>
<p>Your data now lives within this company’s human capital management system. Even if they collect them, very few companies are looking at résumés anymore; they’re looking at the info <a href="https://sea.mashable.com/apps-and-software/23028/this-39-ai-powered-resume-app-could-help-outsmart-application-scanners-online">you type into those tiny boxes</a> to help make comparisons between you, dozens or hundreds of other applicants, and the job requirements. Even if your résumé demonstrates that you are the most qualified applicant, it alone is unlikely to catch the eye of the recruiter, because the recruiter’s attention is elsewhere.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/525995/original/file-20230512-28-a35zgz.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="Six people sit at a table in a brightly lit office with pie charts on the wall and windows in the background." src="https://images.theconversation.com/files/525995/original/file-20230512-28-a35zgz.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/525995/original/file-20230512-28-a35zgz.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/525995/original/file-20230512-28-a35zgz.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/525995/original/file-20230512-28-a35zgz.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/525995/original/file-20230512-28-a35zgz.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/525995/original/file-20230512-28-a35zgz.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/525995/original/file-20230512-28-a35zgz.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">With input from artificial intelligence, measures of employee performance will become more objective and increasingly data-driven.</span>
<span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/business-meeting-in-a-modern-office-royalty-free-image/607477457?phrase=WORKPLACE&adppopup=true">Hinterhaus Productions/DigitalVision via Getty Images</a></span>
</figcaption>
</figure>
<h2>Getting the job</h2>
<p>Let’s say you get the call, you ace the interview and the job is yours. Your information hits another stage within the company’s database, or HCM: active employee. Your performance ratings and other data about your employment will now be tied to your profile, adding more data for the HCM and human resources to monitor and assess.</p>
<p>Enhancements in AI, technology and <a href="https://www.informationweek.com/big-data/the-future-of-hr-tech-how-ai-is-transforming-human-resources">HCMs enable HR</a> to look at employee data on deeper levels. The insights gleaned help identify talented employees who could fill key leadership roles when people quit and guide decisions about who should be promoted. The data can also identify favoritism and bias in hiring and promotion.</p>
<p>As you continue in your role, data on your performance is tracked and analyzed. This may include your performance ratings, supervisor’s feedback, <a href="https://trainingindustry.com/articles/learning-technologies/improving-the-employee-lifecycle-with-ai-enabled-learning-technologies/">professional development activity</a> – or lack thereof. Having this large amount of data about you and others over time now helps HR think about how employees can better support the growth of the organization.</p>
<p>For example, HR may use data to identify how <a href="https://www.shrm.org/ResourcesAndTools/hr-topics/technology/Pages/How-to-Identify-Your-Companys-Flight-Risks.aspx">likely specific employees are to quit</a> and evaluate the impact of that loss. </p>
<p>Platforms that many people already use every day aggregate productivity data from sign-in to signoff. Widely available Microsoft tools including Teams, Outlook and SharePoint can help provide insight to managers via their <a href="https://learn.microsoft.com/en-us/office365/servicedescriptions/workplace-analytics-service-description">workplace analytics tool</a>. The Microsoft productivity score tracks overall usage within the platform. </p>
<p>Even the metrics and <a href="http://essay.utwente.nl/91198/">behaviors defining</a> “good” or “bad” performance may change, relying less on the perception of the manager. As data grows, even <a href="https://time.com/charter/6242075/how-generative-ai-will-change-all-knowledge-work/">the work of professionals</a> like consultants, <a href="https://hbr.org/2016/10/robots-will-replace-doctors-lawyers-and-other-professionals">doctors</a> and marketers will be quantitatively and objectively measured. A 2022 New York Times investigation found that these systems, designed to improve worker productivity and accountability, had the effect of <a href="https://www.nytimes.com/interactive/2022/08/14/business/worker-productivity-tracking.html">damaging morale and instilling fear</a>. </p>
<p>It’s clear that American employees should begin to think about how our data is being used, what story that data is telling, and how it may dictate our futures. </p>
<h2>Optimizing and understanding your career</h2>
<p>Not every company has an HCM or is advanced in using talent data to make decisions. But many companies are becoming savvier and some are incredibly advanced. At a recent Microsoft Viva summit I attended, chief human resources officers from companies like PayPal and Rio Tinto outlined ways they are using these advancements. </p>
<p><a href="https://doi.org/10.1007/s13347-020-00406-7">Some researchers claim</a> that AI could promote equity by removing implicit bias from hiring and promoting, but many more see a danger that AI built by humans will just repackage <a href="https://apo.org.au/node/210071">old issues in a new box</a>. <a href="https://www.imd.org/contentassets/7bcfa11250bc43c994c4975c50f13f8f/tc061-18-print.pdf">Amazon learned this lesson the hard way</a> back in 2018 when a résumé-sorting AI it built had to be abandoned when <a href="https://www.theguardian.com/technology/2018/oct/10/amazon-hiring-ai-gender-bias-recruiting-engine">it favored men for programming roles</a>. </p>
<p>What’s more, the increase of data collection and analysis can leave employees unclear on where they stand while the organization is very clear. It’s best if you understand how AI is changing the workplace and demand transparency from your employer. These are data points that employees should consider asking about during their next review:</p>
<ul>
<li>Do you see me as a <a href="https://doi.org/10.1108/CDI-06-2017-0095">high-potential employee</a>?</li>
<li>How does my performance compare with others’?</li>
<li>Do you see me as a <a href="https://doi.org/10.1177/0892020619881044">successor</a> to your role or others’?</li>
</ul>
<p>Just as you need to master traditional aspects of workplace culture, politics and relationships, you should learn to navigate these platforms, understand how you are being assessed, and take ownership of your career in a new and more data-driven way.</p><img src="https://counter.theconversation.com/content/203318/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>I work as a consultant in corporate settings and use/have used various HCMs and Microsoft tools referenced here in that work.</span></em></p>Software you may already use every day can track your productivity for your employer.Catherine Rymsha, Visiting Lecturer of Management, UMass LowellLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2029212023-05-02T12:13:53Z2023-05-02T12:13:53ZMath teachers hold a bias against girls when the teachers think gender equality has been achieved<figure><img src="https://images.theconversation.com/files/523481/original/file-20230428-22-ygf772.jpg?ixlib=rb-1.1.0&rect=0%2C92%2C7744%2C5143&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Effects of biases can snowball over time.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/shot-of-a-little-girl-doing-maths-on-a-board-in-a-royalty-free-image/1391720645?phrase=girls+math&adppopup=true">PeopleImages via Getty Images</a></span></figcaption></figure><p><em>The <a href="https://theconversation.com/us/topics/research-brief-83231">Research Brief</a> is a short take about interesting academic work.</em> </p>
<h2>The big idea</h2>
<p>Math teachers who believe women no longer face discrimination tend to be biased against girls’ ability in math. This is what we found through an <a href="https://doi.org/10.1186/s40594-023-00420-z">experiment we conducted</a> with over 400 elementary and middle school math teachers across the United States. Our findings were published in a peer-reviewed article that appeared in April 2023 in the International Journal of STEM Education.</p>
<p>For our experiment, we asked teachers to evaluate a set of student solutions to math problems. The teachers didn’t know that gender- and race-specific names, such as Tanisha and Connor, had been randomly assigned to the solutions. We did this so that if they evaluated identical student work differently, it would be because of the gender- and race-specific names they saw, not the differences in student work. The idea was to see if the teachers had any unconscious biases.</p>
<p>After the teachers evaluated the student solutions, we asked a series of questions about their beliefs and experiences. We asked if they felt society had achieved gender equality. We asked them whether they felt anxious about doing math. We asked whether they felt students’ ability in math was fixed or could be improved. We also asked teachers to think about their own experience as math students and to report how frequently they experienced feelings of unequal treatment because of their race or gender.</p>
<p>We then investigated if these beliefs and experiences were related to how they evaluated the math ability of students of different genders or racial groups. </p>
<p>Consistent with our <a href="https://doi.org/10.3102/0013189X19890577">prior work</a>, we found that implicit bias against girls arises in ambiguous situations — in this case, when student solutions were not completely correct.</p>
<p>Further, for teachers who believed that U.S. society had achieved gender equality, they tended to rate a student’s ability higher when they saw a male student name than when they saw a female student name for the same student work.</p>
<h2>Why it matters</h2>
<p>Teachers’ unconscious <a href="https://doi.org/10.1016/j.compedu.2022.104627">gender biases</a> in math classes have been documented repeatedly.</p>
<p>Our study identifies factors that underlie such biases; namely, that biases are stronger among teachers who believe that gender discrimination is not a problem in the United States. Understanding the relationship between teachers’ beliefs and biases can help teacher educators create effective and targeted interventions to remove such biases from classrooms. </p>
<p>Our findings also shed light on potential reasons that <a href="https://doi.org/10.1016/j.lindif.2016.01.002">males tend to have higher confidence in math</a> and <a href="https://doi.org/10.1126/science.aba7377">stick with math-intensive college majors</a> even when they’re not high performers.</p>
<h2>What still isn’t known</h2>
<p>One big remaining question is how to create targeted interventions to help teachers overcome such biases. Evidence suggests that unconscious biases <a href="https://doi.org/10.1177/17456916211057565">come into play in situations where stereotypes might emerge</a>. Further, research suggests that these unconscious biases can be suppressed only when people are aware of them and motivated to restrain them.</p>
<p>Since bias may take on different forms in different fields, a one-time, one-size-fits-all anti-bias training <a href="https://doi.org/10.1037/pspa0000160">may not have a lasting effect</a>. We think it’s worthwhile to investigate if it’s more effective to provide implicit bias training programs that are specific to the areas where bias is revealed.</p><img src="https://counter.theconversation.com/content/202921/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Joseph Cimpian receives funding from the U.S. Department of Education Institute of Education Sciences and the National Science Foundation. </span></em></p><p class="fine-print"><em><span>Ian Thacker and Yasemin Copur-Gencturk do not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.</span></em></p>Teachers judged the same math work differently based on whether the work was associated with male or female names.Yasemin Copur-Gencturk, Associate Professor of Education, University of Southern CaliforniaIan Thacker, Assistant Professor of Educational Psychology, The University of Texas at San AntonioJoseph Cimpian, Professor of Economics and Education Policy, New York UniversityLicensed as Creative Commons – attribution, no derivatives.