tag:theconversation.com,2011:/global/topics/search-37445/articles
Search – The Conversation
2023-03-15T12:24:20Z
tag:theconversation.com,2011:article/200875
2023-03-15T12:24:20Z
2023-03-15T12:24:20Z
AI information retrieval: A search engine researcher explains the promise and peril of letting ChatGPT and its cousins search the web for you
<figure><img src="https://images.theconversation.com/files/515315/original/file-20230314-3889-9mnutz.jpg?ixlib=rb-1.1.0&rect=0%2C0%2C3572%2C2507&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Large language model AI responds to questions but doesn't actually know anything and is prone to making things up.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/big-red-robot-toy-teaches-his-class-royalty-free-image/915921044">Charles Taylor/iStock via Getty Images</a></span></figcaption></figure><p>The prominent model of information access before search engines became the norm – librarians and subject or search experts providing relevant information – was interactive, personalized, transparent and authoritative. Search engines are the primary way most people access information today, but entering a few keywords and getting a list of results ranked by some unknown function is not ideal.</p>
<p>A new generation of artificial intelligence-based information access systems, which includes Microsoft’s <a href="https://www.bing.com/new">Bing/ChatGPT</a>, <a href="https://blog.google/technology/ai/bard-google-ai-search-updates/">Google/Bard</a> and <a href="https://ai.facebook.com/blog/large-language-model-llama-meta-ai/">Meta/LLaMA</a>, is upending the traditional search engine mode of search input and output. These systems are able to take full sentences and even paragraphs as input and generate personalized natural language responses. </p>
<p>At first glance, this might seem like the best of both worlds: personable and custom answers combined with the breadth and depth of knowledge on the internet. But as a researcher who <a href="https://scholar.google.com/citations?user=H4dLAw0AAAAJ&hl=en">studies the search and recommendation systems</a>, I believe the picture is mixed at best.</p>
<p>AI systems like ChatGPT and Bard are built on large language models. A language model is a machine-learning technique that uses a large body of available texts, such as Wikipedia and PubMed articles, to learn patterns. In simple terms, these models figure out what word is likely to come next, given a set of words or a phrase. In doing so, they are able to <a href="https://dx.doi.org/10.2139/ssrn.4358789">generate sentences, paragraphs and even pages</a> that correspond to a query from a user. On March 14, 2023, OpenAI announced the next generation of the technology, GPT-4, which <a href="https://openai.com/research/gpt-4">works with both text and image input</a>, and Microsoft announced that its <a href="https://techcrunch.com/2023/03/14/microsofts-new-bing-was-using-gpt-4-all-along/">conversational Bing is based on GPT-4</a>.</p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/1wzPr4cUoMQ?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
<figcaption><span class="caption">‘60 Minutes’ looked at the good and the bad of ChatGPT.</span></figcaption>
</figure>
<p>Thanks to the training on large bodies of text, fine-tuning and other machine learning-based methods, this type of information retrieval technique works quite effectively. The large language model-based systems generate personalized responses to fulfill information queries. People have found the results so impressive that ChatGPT reached 100 million users in one third of the time it took TikTok to get to that milestone. People have used it to not only find answers but to <a href="https://www.statnews.com/2023/02/13/chatgpt-assisted-diagnosis/">generate diagnoses</a>, <a href="https://www.cnet.com/tech/computing/the-five-best-uses-so-far-for-chatgpts-ai-chatbot/">create dieting plans</a> and <a href="https://www.upstart.com/learn/ai-for-personal-finance/">make investment recommendations</a>.</p>
<h2>Opacity and ‘hallucinations’</h2>
<p>However, there are plenty of downsides. First, consider what is at the heart of a large language model – a mechanism through which it connects the words and presumably their meanings. This produces an output that often seems like an intelligent response, but large language model systems are <a href="https://doi.org/10.1145/3442188.3445922">known to produce almost parroted statements</a> without a real understanding. So, while the generated output from such systems might seem smart, it is merely a reflection of underlying patterns of words the AI has found in an appropriate context. </p>
<p>This limitation makes large language model systems susceptible to making up or <a href="https://www.datanami.com/2023/01/17/hallucinations-plagiarism-and-chatgpt/">“hallucinating” answers</a>. The systems are also not smart enough to understand the incorrect premise of a question and answer faulty questions anyway. For example, when asked which U.S. president’s face is on the $100 bill, ChatGPT answers Benjamin Franklin without realizing that Franklin was never president and that the premise that the $100 bill has a picture of a U.S. president is incorrect. </p>
<p>The problem is that even when these systems are wrong only 10% of the time, you don’t know which 10%. People also don’t have the ability to quickly validate the systems’ responses. That’s because these systems lack transparency – they don’t reveal what data they are trained on, what sources they have used to come up with answers or how those responses are generated. </p>
<p>For example, you could ask ChatGPT to write a technical report with citations. But often it <a href="https://blogs.library.duke.edu/blog/2023/03/09/chatgpt-and-fake-citations/">makes up these citations</a> – “hallucinating” the titles of scholarly papers as well as the authors. The systems also don’t validate the accuracy of their responses. This leaves the validation up to the user, and users may not have the motivation or skills to do so or even recognize the need to check an AI’s responses.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/514788/original/file-20230311-4402-j31o82.png?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="screen capture of text" src="https://images.theconversation.com/files/514788/original/file-20230311-4402-j31o82.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/514788/original/file-20230311-4402-j31o82.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=141&fit=crop&dpr=1 600w, https://images.theconversation.com/files/514788/original/file-20230311-4402-j31o82.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=141&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/514788/original/file-20230311-4402-j31o82.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=141&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/514788/original/file-20230311-4402-j31o82.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=177&fit=crop&dpr=1 754w, https://images.theconversation.com/files/514788/original/file-20230311-4402-j31o82.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=177&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/514788/original/file-20230311-4402-j31o82.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=177&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 doesn’t know when a question doesn’t make sense, because it doesn’t know any facts.</span>
<span class="attribution"><span class="source">Screen capture by Chirag Shah</span></span>
</figcaption>
</figure>
<h2>Stealing content – and traffic</h2>
<p>While lack of transparency can be harmful to the users, it is also unfair to the authors, artists and creators of the original content from whom the systems have learned, because the systems do not reveal their sources or provide sufficient attribution. In most cases, creators are <a href="https://medium.com/inspired-writer/yes-chatgpt-can-hurt-writers-but-not-the-way-you-think-f5118e0307db">not compensated or credited</a> or given the opportunity to give their consent.</p>
<p>There is an economic angle to this as well. In a typical search engine environment, the results are shown with the links to the sources. This not only allows the user to verify the answers and provides the attributions to those sources, it also <a href="https://searchengineland.com/organic-search-responsible-for-53-of-all-site-traffic-paid-15-study-322298">generates traffic for those sites</a>. Many of these sources rely on this traffic for their revenue. Because the large language model systems produce direct answers but not the sources they drew from, I believe that those sites are likely to see their revenue streams diminish.</p>
<h2>Taking away learning and serendipity</h2>
<p>Finally, this new way of accessing information also can disempower people and takes away their chance to learn. A typical search process allows users to explore the range of possibilities for their information needs, often triggering them to adjust what they’re looking for. It also affords them an <a href="https://doi.org/10.1145/3176349.3176386">opportunity to learn</a> what is out there and how various pieces of information connect to accomplish their tasks. And it allows for <a href="https://doi.org/10.1108/00220410310472518">accidental encounters or serendipity</a>.</p>
<p>These are very important aspects of search, but when a system produces the results without showing its sources or guiding the user through a process, it robs them of these possibilities.</p>
<p>Large language models are a great leap forward for information access, providing people with a way to have natural language-based interactions, produce personalized responses and discover answers and patterns that are often difficult for an average user to come up with. But they have severe limitations due to the way they learn and construct responses. Their answers may be <a href="https://www.fastcompany.com/90820090/the-internet-loves-chatgpt-but-theres-a-dark-side-to-the-tech">wrong, toxic or biased</a>.</p>
<p>While other information access systems can suffer from these issues, too, large language model AI systems also lack transparency. Worse, their natural language responses can help fuel a <a href="https://nymag.com/intelligencer/article/ai-artificial-intelligence-chatbots-emily-m-bender.html">false sense of trust and authoritativeness</a> that can be dangerous for uninformed users.</p><img src="https://counter.theconversation.com/content/200875/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Chirag Shah receives funding from National Science Foundation (NSF). </span></em></p>
Searching the web with ChatGPT is like talking to an expert – if you’re OK getting a mix of fact and fiction. But even if it were error-free, searching this way comes with hidden costs.
Chirag Shah, Professor of Information Science, University of Washington
Licensed as Creative Commons – attribution, no derivatives.
tag:theconversation.com,2011:article/189271
2022-08-23T23:40:45Z
2022-08-23T23:40:45Z
Trump’s Mar-a-Lago lawsuit spotlights how difficult search warrants are to challenge – by a criminal suspect or an ex-president – until charges are brought
<figure><img src="https://images.theconversation.com/files/480644/original/file-20220823-25-62ver2.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Mar-a-Lago is shown on Aug. 16, 2022, a week after the FBI's raid. </span> <span class="attribution"><a class="source" href="https://media.gettyimages.com/photos/maralago-is-seen-august-16-2022-a-week-after-the-fbi-raided-the-home-picture-id1242585509?s=2048x2048">Nathan Posner/Anadolu Agency via Getty Images</a></span></figcaption></figure><p>Some observers say that the lawsuit filed by former President Donald Trump on <a href="https://www.axios.com/2022/08/22/donald-trump-fbi-search-lawsuit">Aug. 22, 2022,</a> challenging the <a href="https://theconversation.com/why-searching-an-ex-presidents-estate-is-not-easily-done-4-important-things-to-know-about-the-fbis-search-of-mar-a-lago-188438">FBI’s recent search</a> of his Mar-a-Lago estate is “filled with <a href="https://www.nytimes.com/2022/08/22/us/politics/trump-warrant-affidavit-reinhart.html">bombastic complaints”</a> and will <a href="https://twitter.com/PoliticusSarah/status/1561844965269012482">“blow up in his face.”</a>.</p>
<p>I am <a href="http://www.clarkcunningham.org/">a legal scholar</a> who is an expert on the various<a href="https://news.gsu.edu/2022/07/06/cunningham-legal-voice-for-jan-6-hearings-2020-presidential-election-investigation/"> Trump investigations</a> and the <a href="https://www.yalelawjournal.org/forum/apple-and-the-american-revolution-remembering-why-we-have-the-fourth-amendment-1">constitutional protections</a> against <a href="https://theconversation.com/in-getting-new-clinton-emails-did-the-fbi-violate-the-constitution-67906">wrongful searches</a>. </p>
<p>I think it is important to recognize that <a href="https://storage.courtlistener.com/recap/gov.uscourts.flsd.618763/gov.uscourts.flsd.618763.1.0.pdf">Trump’s lawsuit</a> raises a very serious point: Current federal law does not provide good procedures to protect the rights of people subjected to a search warrant. </p>
<h2>Federal law’s limits on searches</h2>
<p>The Constitution protects “<a href="https://constitution.congress.gov/constitution/amendment-4/">the right of the people to be secure in their houses and papers</a>” and requires that search warrants must “<a href="https://constitution.congress.gov/constitution/amendment-4/">particularly describe</a>” the place to be searched and the things to be seized. </p>
<p><a href="https://storage.courtlistener.com/recap/gov.uscourts.flsd.618763/gov.uscourts.flsd.618763.1.0.pdf">Trump’s lawsuit </a> asserts that these Constitutional protections were violated both by the broad language of the search warrant and the way it has been carried out by the FBI. </p>
<p><a href="https://www.bbc.com/news/world-us-canada-62640448">Trump’s lawsuit</a> asks for a judge to halt the FBI’s review of what it seized from Mar-a-Lago and the appointment of an independent judicial officer to conduct the review instead.</p>
<p>Trump also asks for a more detailed receipt of what the FBI took, and for the agency to return all of the items not properly seized. </p>
<p>Granting these requests might both be fair to Trump and also in the public interest, by bolstering public confidence in the handling of the search, which has been <a href="https://www.washingtonpost.com/national-security/2022/08/20/fbi-comes-under-threat-its-leader-tries-stay-out-view/">criticized by Trump and his supporters </a>as politically motivated, intrusive and overbroad.</p>
<p>However, as shown by <a href="https://www.politico.com/news/2022/06/27/eastman-phone-seized-jan-6-00042680">a similar lawsuit </a>recently filed by one of Trump’s former lawyers, John Eastman, people who are subjected to a federal search have limited ability to challenge its legality.</p>
<p>Eastman’s <a href="https://storage.courtlistener.com/recap/gov.uscourts.nmd.476086/gov.uscourts.nmd.476086.1.0_1.pdf">cellphone was taken</a> by federal agents as he was leaving a restaurant in <a href="https://www.washingtonpost.com/national-security/2022/06/27/eastman-phone-seized-fbi-jan6/">June 2022</a>, amid a <a href="https://www.theguardian.com/us-news/2022/jul/16/john-eastman-cell-phone-seized-january-6">federal investigation </a>into his alleged attempt to overturn the 2020 election results. </p>
<p>The <a href="https://storage.courtlistener.com/recap/gov.uscourts.nmd.476086/gov.uscourts.nmd.476086.1.1.pdf">search warrant</a> authorized seizing “any and all electronic or digital devices and all information in such devices” without identifying what crime was being investigated. Eastman <a href="https://www.cbsnews.com/news/trump-attorney-john-eastman-says-fbi-agents-seized-his-phone-according-to-new-lawsuit/">challenged the warrant</a> on the same grounds as Trump’s lawsuit, claiming it authorized an overbroad search of everything stored on his phone. </p>
<p>Eastman tried to get a federal court to halt FBI examination of his phone by invoking a Federal Rule of Criminal Procedure, or Rule 41, which says “<a href="https://www.law.cornell.edu/rules/frcrmp/rule_41">a person aggrieved by an unlawful search and seizure</a>” may seek return of the seized property. In response, the government told the court that Eastman can only use Rule 41 to recover property for which he can show an urgent need. Eastman cannot use Rule 41 to challenge the constitutionality of the FBI seizing his phone or to prevent the law enforcement agency from reading attorney-client communications stored on the phone, the government said. </p>
<p>Trump’s lawsuit faces the same problem. Apart from Rule 41, there is currently no clear way under federal law to challenge the validity of a search unless and until criminal charges are filed. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/480647/original/file-20220823-13-vdkxeh.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="People hold flags and signs that say 'The FBI is corrupt' and 'crimes happen here' on a street corner on a sunny day." src="https://images.theconversation.com/files/480647/original/file-20220823-13-vdkxeh.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/480647/original/file-20220823-13-vdkxeh.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/480647/original/file-20220823-13-vdkxeh.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/480647/original/file-20220823-13-vdkxeh.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/480647/original/file-20220823-13-vdkxeh.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/480647/original/file-20220823-13-vdkxeh.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/480647/original/file-20220823-13-vdkxeh.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">Demonstrators protest the FBI’s recent raid of Mar-a-Lago outside the agency’s Chelsea, Mass., building on Aug. 21, 2022.</span>
<span class="attribution"><a class="source" href="https://media.gettyimages.com/photos/demonstrators-wave-flags-and-hold-signs-as-they-protest-the-recent-picture-id1242639842?s=2048x2048">Joseph Prezioso/AFP via Getty Images</a></span>
</figcaption>
</figure>
<h2>An often overlooked omission</h2>
<p>This loophole in federal law – and American law generally – means that court cases about searches are almost always in the context of criminal prosecution. </p>
<p>Because those making court arguments for rights against improper searches are usually accused or convicted criminals, the general public has paid little attention to the fact that search <a href="https://theconversation.com/feds-we-can-read-all-your-email-and-youll-never-know-65620">warrant procedures are an exception</a> to a fundamental principle of American law, which is that people <a href="https://theconversation.com/restoring-transparency-and-fairness-to-the-fbi-investigation-of-clinton-emails-67967">have the right to participate</a> in judicial proceedings regarding their rights. </p>
<p>As pointed out by the government <a href="https://www.courtlistener.com/docket/64872441/united-states-v-sealed-search-warrant/">in opposing the unsealing of the FBI affidavit </a>used to obtain the Mar-a-Lago warrant, federal courts consistently allow investigative records to be sealed from both the subjects of investigation and the public. A judge’s initial decision to issue a search warrant is almost always based only on a one-sided presentation by the government. </p>
<p>Not only do subjects of a requested warrant have no chance to present their side to the judge, but they do not even know about the warrant process until the government is at the door, warrant in hand. And then, <a href="https://storage.courtlistener.com/recap/gov.uscourts.flsd.618763/gov.uscourts.flsd.618763.1.0.pdf">as Trump’s lawsuit complains</a>, this secrecy continues after the search is conducted as the government reviews what it seized.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/480648/original/file-20220823-23-rz4lgs.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="A large brown residential appearing building is shown on a day with dark clouds." src="https://images.theconversation.com/files/480648/original/file-20220823-23-rz4lgs.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/480648/original/file-20220823-23-rz4lgs.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/480648/original/file-20220823-23-rz4lgs.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/480648/original/file-20220823-23-rz4lgs.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/480648/original/file-20220823-23-rz4lgs.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/480648/original/file-20220823-23-rz4lgs.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/480648/original/file-20220823-23-rz4lgs.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">Mar-a-Lago is seen on Aug. 16, 2022.</span>
<span class="attribution"><a class="source" href="https://media.gettyimages.com/photos/maralago-is-seen-august-16-2022-a-week-after-the-fbi-raided-the-home-picture-id1242585654?s=2048x2048">Nathan Posner/Anadolu Agency via Getty Images</a></span>
</figcaption>
</figure>
<h2>Trump’s request for a ‘special master’</h2>
<p>The FBI is now <a href="https://www.politico.com/news/2022/08/22/trump-files-suit-special-master-mar-a-lago-search-00053196">using a “taint team”</a> for the Mar-a-Lago search and seizure of classified documents. This is a special group of agents designated to do an initial review of seized materials. </p>
<p>These screening officers then decide what materials can be turned over for further review by FBI agents doing the actual criminal investigation. In Trump’s case, however, the standards for such review are not public and nothing in the court record indicates review standards have been submitted for court approval. </p>
<p>Even if review criteria are clearly defined, the practice of using FBI agents for screening has been <a href="https://casetext.com/case/united-states-v-seal-in-re-search-warrant-issued-june">criticized by some courts</a> as providing insufficient protection against improper use of items seized during a search.</p>
<p>For example, <a href="https://law.justia.com/cases/federal/appellate-courts/F3/454/511/489658/">one federal appellate court</a> described taint teams as putting the fox in charge of guarding the hen house. Even on the taint team, FBI agents may still have a “<a href="https://casetext.com/case/united-states-v-seal-in-re-search-warrant-issued-june">prosecutorial interest</a>” that could lead them to hand over documents to investigators that should be protected from government view. </p>
<p>This kind of action, whether because of <a href="https://law.justia.com/cases/federal/appellate-courts/F3/454/511/489658/">malice, neglect or simply an honest mistake</a>, can take place before the subjects of the search have an opportunity to seek court protections for their documents.</p>
<p>That court did what the Trump lawsuit is now requesting – <a href="https://www.lexology.com/library/detail.aspx?g=74560be8-27fc-4ff3-bdaf-f02e56138634">it ordered</a> that a temporary judicial officer, called a “<a href="https://www.law.cornell.edu/wex/special_master">special master</a>,” take over the initial review, to exclude documents that should not be seen by the government. </p>
<p>There are other instances of commissioning a person to do such a job. In 2018, for example, when federal agents <a href="https://www.nytimes.com/2018/04/26/nyregion/michael-cohen-investigation-special-master.html">executed search warrants</a> against Trump’s former personal attorney, Michael D. Cohen, a retired federal judge was appointed as a special master to screen everything that was seized before it could be turned over to prosecutors. </p>
<p>The justification for placing an independent judicial officer between the FBI and the trove of documents seized at Mar-a-Lago is underscored by the broad way the <a href="https://int.nyt.com/data/documenttools/mar-a-lago-search-warrant-and-inventory/6478c5980764438f/full.pdf">search warrant</a> was written. </p>
<p>The warrant not only authorized the FBI to seize classified documents, but it also allowed the FBI to seize “any other containers/boxes” that were “stored or found together” with boxes containing classified documents. </p>
<p>This means it is possible that some of the 26 boxes listed on <a href="https://int.nyt.com/data/documenttools/mar-a-lago-search-warrant-and-inventory/6478c5980764438f/full.pdf">the FBI’s property receipt</a> were seized not because they contained evidence of a crime but simply because they were stored in the same location as classified documents. Without something like the protections of a special master procedure, FBI agents could end up reading thousands of pages taken from Trump’s home that have no relevance to the suspected crimes listed in the warrant.</p>
<p>Perhaps now that the one-sided nature of search warrant procedures is being challenged by a former president, this problem will get new attention.</p><img src="https://counter.theconversation.com/content/189271/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Clark D. Cunningham 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>
Trump’s lawsuit against the FBI has been criticized as baseless. But it spotlights a loophole in federal law that doesn’t protect people’s rights when they are subjected to a search warrant.
Clark D. Cunningham, W. Lee Burge Chair in Law & Ethics; Director, National Institute for Teaching Ethics & Professionalism, Georgia State University
Licensed as Creative Commons – attribution, no derivatives.
tag:theconversation.com,2011:article/188438
2022-08-09T18:12:53Z
2022-08-09T18:12:53Z
Why searching an ex-president’s estate is not easily done – 4 important things to know about the FBI’s search of Mar-a-Lago
<figure><img src="https://images.theconversation.com/files/478330/original/file-20220809-15076-ffr1dv.jpg?ixlib=rb-1.1.0&rect=0%2C100%2C4496%2C2887&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Palm Beach police officers stand near the Florida home of former President Donald Trump on Aug. 8, 2022. </span> <span class="attribution"><a class="source" href="https://media.gettyimages.com/photos/palm-beach-police-officers-keep-watch-near-the-home-of-former-donald-picture-id1242395984?s=2048x2048">Eva Marie Uzcategui/Getty Images </a></span></figcaption></figure><p><em>The FBI’s raid of former President Donald Trump’s estate on Aug. 8, 2022, caught Trump by surprise – and prompted immediate speculation about exactly <a href="https://thehill.com/homenews/3594254-heres-what-we-know-about-the-fbi-search-of-trumps-mar-a-lago/">why and how</a> the law enforcement agency secured a search warrant.</em></p>
<p><em>“My beautiful home, Mar-A-Lago in Palm Beach, Florida, is currently under siege, raided, and occupied by a large group of FBI agents. … They even broke into my safe!” <a href="https://twitter.com/kyledcheney/status/1556775612920074240">Trump said in a statement</a> released through his political action committee, Save America.</em></p>
<p><em>Trump brought <a href="https://www.washingtonpost.com/politics/2022/02/12/trump-15-boxes/">15 boxes of classified materials</a> with him to Mar-a-Lago when he left the White House, and delayed returning the materials to National Archives officials for months.</em></p>
<p><em>The FBI and the Department of Justice have not commented on the raid, but the Justice Department is <a href="https://www.npr.org/2022/04/07/1091431136/justice-department-investigating-trumps-possible-mishandling-of-government-secre">known to be investigating</a> how Trump possibly mishandled government secrets. Trump is also facing other potential charges from the <a href="https://www.nytimes.com/2022/07/19/us/georgia-trump-electors.html">state of Georgia</a> stemming from his alleged interference with the 2020 elections.</em></p>
<p><em>Georgia State University legal scholar <a href="http://www.clarkcunningham.org/">Clark D. Cunningham</a>, an expert on <a href="https://www.yalelawjournal.org/forum/apple-and-the-american-revolution-remembering-why-we-have-the-fourth-amendment-1">search warrants</a> and the criminal <a href="https://news.gsu.edu/2022/07/06/cunningham-legal-voice-for-jan-6-hearings-2020-presidential-election-investigation/">investigations of interference</a> in the 2020 election, explains what could have led to the raid and what the raid tells us about the state of the federal investigation into Trump’s activities.</em></p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/478308/original/file-20220809-14165-ffr1dv.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="An older white man is shown seated at a desk, gesticulating with his mouth open, in an ornate-looking room. In front of him is a group of reporters and camera people with equipment." src="https://images.theconversation.com/files/478308/original/file-20220809-14165-ffr1dv.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/478308/original/file-20220809-14165-ffr1dv.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/478308/original/file-20220809-14165-ffr1dv.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/478308/original/file-20220809-14165-ffr1dv.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/478308/original/file-20220809-14165-ffr1dv.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/478308/original/file-20220809-14165-ffr1dv.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/478308/original/file-20220809-14165-ffr1dv.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">Former President Donald Trump speaks to the press from his Mar-a-Lago resort in Palm Beach, Fla., in 2018.</span>
<span class="attribution"><a class="source" href="https://media.gettyimages.com/photos/president-donald-trump-speaks-to-the-press-after-talking-to-members-picture-id1064313534?s=2048x2048">Mandel Ngan/AFP via Getty Images</a></span>
</figcaption>
</figure>
<h2>1. There are legal hurdles to getting a search warrant</h2>
<p>The <a href="https://constitution.congress.gov/constitution/amendment-4/">U.S. Constitution requires</a> that all search warrants “particularly describe the place to be searched and the … things to be seized.” </p>
<p>This requirement can be traced in part to a <a href="https://www.yalelawjournal.org/forum/apple-and-the-american-revolution-remembering-why-we-have-the-fourth-amendment-1">famous British case from the 1760s</a> when agents of King George III searched the house of <a href="https://www.britannica.com/biography/John-Wilkes">John Wilkes</a>, an opposition member of Parliament, for incriminating papers. The warrant they used was condemned by the courts as a “general warrant” because it did not specifically name Wilkes, his house or the seized papers. </p>
<p>Courts and commentators also criticized the Wilkes warrant because it was based on mere suspicion. The U.S. founders looked to the Wilkes warrant as an example of what the Constitution should prevent and added <a href="https://constitution.congress.gov/constitution/amendment-4/">the Fourth Amendment</a> – requiring that search warrants only be issued “upon probable cause, supported by Oath.” </p>
<p>Criminal <a href="https://www.law.cornell.edu/rules/frcrmp/rule_41">procedure laws</a> help enforce these constitutional requirements by requiring search warrants to particularly describe “evidence of a crime … or other items illegally possessed.” </p>
<p>Only judges can issue search warrants, and they must find, <a href="https://www.law.cornell.edu/rules/frcrmp/rule_41">based on sworn testimony</a>, that there is probable cause that such evidence or items will be found in the location described in the warrant.</p>
<p>This means that a judge must have found that there was probable cause that either a crime had been committed, or that Trump was illegally possessing items taken from the White House. The FBI’s request for a search warrant might also have indicated concern that these documents would either be destroyed or moved off of the premises.</p>
<h2>2. There are also potential policy hurdles</h2>
<p>In February 2020, then-Attorney General William Barr <a href="https://www.nytimes.com/2020/02/05/us/politics/barr-2020-investigations.html">announced new restrictions </a> that require the FBI and other law enforcement agencies to <a href="https://www.nytimes.com/2020/02/05/us/politics/barr-2020-investigations.html">get permission</a> from the Attorney General before investigating presidential candidates or their staff. </p>
<p>Barr’s successor, Attorney General Merrick Garland, has <a href="https://www.nytimes.com/2022/07/28/us/politics/trump-garland-investigation.html">kept this policy in place</a> – keeping in line with general <a href="https://www.justice.gov/about">Justice Department guidelines</a> that try to prevent politically charged investigations. </p>
<p>This means that this search would not have taken place without Garland’s approval. Given the generally strong tradition of <a href="https://www.lawfareblog.com/independence-and-accountability-department-justice">political independence</a> at the Justice Department, it is not surprising that President Joe <a href="https://www.newsweek.com/did-biden-know-about-fbi-search-trumps-mar-lago-what-we-know-1732190">Biden and</a> his aides <a href="https://www.vox.com/policy-and-politics/2022/8/9/23297734/donald-trump-mar-a-lago-fbi-raid">were not informed</a> in advance of the raid and found out on Twitter. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/478313/original/file-20220809-18-tvikwd.jpg?ixlib=rb-1.1.0&rect=100%2C18%2C4055%2C2747&q=45&auto=format&w=1000&fit=clip"><img alt="A police officer leans against a police car while a woman walks past. Behind them are large white gates, shining blue and red because of the police lights." src="https://images.theconversation.com/files/478313/original/file-20220809-18-tvikwd.jpg?ixlib=rb-1.1.0&rect=100%2C18%2C4055%2C2747&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/478313/original/file-20220809-18-tvikwd.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=399&fit=crop&dpr=1 600w, https://images.theconversation.com/files/478313/original/file-20220809-18-tvikwd.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=399&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/478313/original/file-20220809-18-tvikwd.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=399&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/478313/original/file-20220809-18-tvikwd.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=502&fit=crop&dpr=1 754w, https://images.theconversation.com/files/478313/original/file-20220809-18-tvikwd.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=502&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/478313/original/file-20220809-18-tvikwd.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=502&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">Former President Donald Trump was not at Mar-a-Lago when the FBI searched the premises on Aug. 8, 2022.</span>
<span class="attribution"><a class="source" href="https://media.gettyimages.com/photos/police-car-is-seen-outside-former-us-president-donald-trumps-in-picture-id1242395292?s=2048x2048">Giorgio Viera/AFP via Getty Images</a></span>
</figcaption>
</figure>
<h2>3. The FBI might have found more than it was looking for</h2>
<p>The Supreme Court ruled in a <a href="https://supreme.justia.com/cases/federal/us/496/128/">1990 case</a> that police executing a warrant that authorized searching for the proceeds of a robbery could also lawfully seize weapons that were in plain view. </p>
<p>Assuming that the FBI’s warrant authorized only searching for classified documents taken from the White House, if the FBI found “in plain view” other evidence of crimes related to the 2020 election or <a href="https://www.cnn.com/2022/07/10/politics/jan-6-us-capitol-riot-timeline/index.html">Jan. 6, 2021, Capitol insurrection</a>, they likely could have taken that, as well. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/478310/original/file-20220809-18-bybrig.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="A few people - one of them yelling - are shown with Trump flags and American flags on a dark evening on the street." src="https://images.theconversation.com/files/478310/original/file-20220809-18-bybrig.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/478310/original/file-20220809-18-bybrig.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/478310/original/file-20220809-18-bybrig.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/478310/original/file-20220809-18-bybrig.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/478310/original/file-20220809-18-bybrig.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/478310/original/file-20220809-18-bybrig.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/478310/original/file-20220809-18-bybrig.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">Supporters of former President Donald Trump protest outside his Mar-a-Lago home following the FBI’s raid on Aug. 8, 2022.</span>
<span class="attribution"><a class="source" href="https://media.gettyimages.com/photos/supporters-of-former-president-donald-trump-shout-as-kamrel-eppinger-picture-id1242396102?s=2048x2048">Eva Marie Uzcategui/Getty Images</a></span>
</figcaption>
</figure>
<h2>4. There may be a connection with Trump’s possible election interference</h2>
<p>A federal <a href="https://www.nytimes.com/2022/05/12/us/politics/justice-department-trump-classified.html">grand jury, requested by the Justice Department, has been investigating</a> the presence of potentially classified documents at Mar-a-Lago since at least early May 2022. It seems likely that something has happened recently to cause this urgent search. One possibility is that the search warrant was issued based on information gathered in <a href="https://www.theguardian.com/us-news/2022/aug/08/donald-trump-fbi-raid-explainer">one or more of the criminal investigations</a> involving 2020 election interference. </p>
<p>In particular, the Department of Justice on July 12, 2022, obtained a <a href="https://int.nyt.com/data/documenttools/filing-by-thomas-windom-in-u-s/c8958e56f1860a88/full.pdf">warrant to search the cellphone</a> of <a href="https://www.pbs.org/newshour/politics/who-is-john-eastman-and-why-is-he-important-to-the-jan-6-hearings">John Eastman</a>, Trump’s former lawyer. As hearings by the Jan. 6 House committee have revealed, Eastman was a <a href="https://www.npr.org/2022/06/17/1105600072/who-is-john-eastman-the-trump-lawyer-at-the-center-of-the-jan-6-investigation">primary architect of the plan</a> to block Congress from certifying Biden’s victory.</p>
<p>There seems little doubt that the Justice Department had compelling, perhaps overwhelming, legal justifications for conducting this unprecedented search of a former president’s home. However, the secrecy required for Justice Department investigations and grand jury proceedings means that the country will have to be patient – the justifications for the search may become public only if and when criminal charges are filed.</p><img src="https://counter.theconversation.com/content/188438/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Clark D. Cunningham 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>
There’s a high bar for a federal judge to grant a search warrant, indicating there is probable cause that Trump committed a crime by holding classified documents at Mar-a-Lago.
Clark D. Cunningham, W. Lee Burge Chair in Law & Ethics; Director, National Institute for Teaching Ethics & Professionalism, Georgia State University
Licensed as Creative Commons – attribution, no derivatives.
tag:theconversation.com,2011:article/137756
2020-05-12T12:35:06Z
2020-05-12T12:35:06Z
AI tool searches thousands of scientific papers to guide researchers to coronavirus insights
<figure><img src="https://images.theconversation.com/files/332869/original/file-20200505-83757-1nuyewy.jpg?ixlib=rb-1.1.0&rect=0%2C0%2C8000%2C4491&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Artificial intelligence can do what humans can't – connect the dots across the majority of coronavirus research.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/coronavirus-royalty-free-image/1215382103?adppopup=true">baranozdemir/E+ via Getty Images</a></span></figcaption></figure><p><em>The Research Brief is a short take about interesting academic work.</em></p>
<h2>The big idea</h2>
<p>The scientific community worldwide has mobilized with unprecedented speed to tackle the COVID-19 pandemic, and the emerging research output is staggering. Every day, <a href="https://covid19primer.com/dashboard">hundreds of scientific papers about COVID-19 come out</a>, in both traditional journals and non-peer-reviewed preprints. There’s already far more than any human could possibly keep up with, and more research is constantly emerging.</p>
<p>And it’s not just new research. We estimate that there are as many as 500,000 papers relevant to COVID-19 that were published before the outbreak, including papers related to the outbreaks of SARS in 2002 and MERS in 2012. Any one of these might contain the key information that leads to effective treatment or a vaccine for COVID-19.</p>
<p>Traditional methods of searching through the research literature just don’t cut it anymore. This is why <a href="https://www.covidscholar.org/about">we and our colleagues</a> at Lawrence Berkeley National Lab are using the latest artificial intelligence techniques to build <a href="https://www.covidscholar.org/">COVIDScholar</a>, a search engine dedicated to COVID-19. COVIDScholar includes tools that pick up subtle clues like similar drugs or research methodologies to recommend relevant research to scientists. AI can’t replace scientists, but it can help them gain new insights from more papers than they could read in a lifetime.</p>
<figure class="align-center ">
<img alt="" src="https://images.theconversation.com/files/333182/original/file-20200506-49538-ey2oya.JPG?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/333182/original/file-20200506-49538-ey2oya.JPG?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=395&fit=crop&dpr=1 600w, https://images.theconversation.com/files/333182/original/file-20200506-49538-ey2oya.JPG?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=395&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/333182/original/file-20200506-49538-ey2oya.JPG?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=395&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/333182/original/file-20200506-49538-ey2oya.JPG?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=496&fit=crop&dpr=1 754w, https://images.theconversation.com/files/333182/original/file-20200506-49538-ey2oya.JPG?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=496&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/333182/original/file-20200506-49538-ey2oya.JPG?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=496&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">COVIDScholar is a search engine with machine learning algorithms under the hood.</span>
<span class="attribution"><span class="source">Screen capture by The Conversation</span></span>
</figcaption>
</figure>
<h2>Why it matters</h2>
<p>When it comes to finding effective treatments for COVID-19, time is of the essence. Scientists spend <a href="https://doi.org/10.1371/journal.pone.0189753">23% of their time searching for and reading papers</a>. Every second our search tools can save them is more time to spend making discoveries in the lab and analyzing data.</p>
<p>AI can do more than just save scientists time. Our group’s previous work showed that AI <a href="https://doi.org/10.1038/s41586-019-1335-8">can capture latent scientific knowledge</a> from text, making connections that humans missed. There, we showed that AI was able to suggest new, cutting-edge functional materials years before their discovery by humans. The information was there all along, but it took combining information from hundreds of thousands of papers to find it.</p>
<p>We are now applying the same techniques to COVID-19, to find existing drugs that could be repurposed, genetic links that might help develop a vaccine or effective treatment regimens. We’re also starting to build in new innovations, like using molecular structures to help find which drugs are similar to each other, including those that are similar in unexpected ways.</p>
<h2>How we do this work</h2>
<p>The most important part of our work is the data. We’ve built web scrapers that collect new papers as they’re published from a wide variety of sources, making them available on our website within 15 minutes of their appearance online. We also clean the data, fixing mistakes in formatting and comparing the same paper from multiple sources to find the best version. Our machine learning algorithms then go to work on the paper, tagging it with subject categories and marking work important to COVID-19.</p>
<figure class="align-center ">
<img alt="" src="https://images.theconversation.com/files/333116/original/file-20200506-49550-10it6ws.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/333116/original/file-20200506-49550-10it6ws.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=181&fit=crop&dpr=1 600w, https://images.theconversation.com/files/333116/original/file-20200506-49550-10it6ws.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=181&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/333116/original/file-20200506-49550-10it6ws.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=181&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/333116/original/file-20200506-49550-10it6ws.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=228&fit=crop&dpr=1 754w, https://images.theconversation.com/files/333116/original/file-20200506-49550-10it6ws.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=228&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/333116/original/file-20200506-49550-10it6ws.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=228&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">COVIDScholar labels and categorizes about 250 journal papers a day to help researchers make connections they might otherwise miss.</span>
<span class="attribution"><span class="source">Kevin Cruse and Haoyan Huo</span>, <a class="license" href="http://creativecommons.org/licenses/by-nd/4.0/">CC BY-ND</a></span>
</figcaption>
</figure>
<p>We’re also continuously seeking out experts in new areas. Their input and annotation of data is what allows us to train new AI models.</p>
<h2>What’s next</h2>
<p>So far, we have assembled a collection of over 60,000 papers on COVID-19, and we’re expanding the collection daily. We’ve also built search tools that group research into categories, suggest related research and allow users to find papers that connect different concepts, such as papers that connect a specific drug to the diseases it’s been used to treat in the past. We’re now building AI algorithms that allow researchers to plug search results into quantitative models for studying topics like protein interactions. We’re also starting to dig through the past literature to find hidden gems.</p>
<p>We hope that very soon, researchers using COVIDScholar will start to identify relationships that they might never have imagined, bringing us closer to treatments and a remedy for COVID-19.</p>
<p>[<em>You’re too busy to read everything. We get it. That’s why we’ve got a weekly newsletter.</em> <a href="https://theconversation.com/us/newsletters/weekly-highlights-61?utm_source=TCUS&utm_medium=inline-link&utm_campaign=newsletter-text&utm_content=weeklybusy">Sign up for good Sunday reading.</a> ]</p><img src="https://counter.theconversation.com/content/137756/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Amalie Trewartha receives funding from the Department of Energy and has previously received funding from the Toyota Research Institute.</span></em></p><p class="fine-print"><em><span>John Dagdelen works for Lawrence Berkeley National Laboratory. He receives funding from Berkeley Lab and previously from the Toyota Research Institute. </span></em></p>
The scientific community is churning out vast quantities of research about the coronavirus pandemic – far too much for researchers to absorb. An AI system aims to do the heavy lifting for them.
Amalie Trewartha, Post Doctoral Fellow, University of California, Berkeley
John Dagdelen, Graduate Student Researcher, Persson Group, University of California, Berkeley
Licensed as Creative Commons – attribution, no derivatives.
tag:theconversation.com,2011:article/110664
2019-01-29T10:14:21Z
2019-01-29T10:14:21Z
MH370: New underwater sound wave analysis suggests alternative travel route and new impact locations
<figure><img src="https://images.theconversation.com/files/255961/original/file-20190128-108348-ythlpc.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-illustration/kuala-lumpur-malaysia-circa-2017-inflight-604236992?src=RWPG0diNk8LwsmIJQmIlwA-2-2">NextNewMedia/Shutterstock</a></span></figcaption></figure><p>Motivated by a desire to help find Malaysia Airlines flight MH370, which is believed to have <a href="https://www.bbc.co.uk/news/world-asia-26716572">crashed into the southern Indian Ocean</a> in March 2014, we proposed a way of working out where objects hit the surface of the ocean <a href="https://theconversation.com/new-tool-to-track-underwater-acoustic-waves-could-find-mh370-86002">using underwater acoustic waves</a>. Unfortunately this didn’t lead to finding the plane. However, our research into these waves has moved on since we first proposed the idea in 2017, and we have now been able to identify two locations where the aeroplane could have impacted with the ocean, as well as an alternative route that the plane may have taken.</p>
<p>When you drop a pebble in a lake, water waves are generated from the location of the impact, while sound waves create the splashing noise you hear. Another type of wave is generated inside the water too: hydroacoustic. Similar to a sound wave, hydroacoustic waves move much faster through the denser water than they would through air – 1,500 metres per second (m/s) compared to 340m/s.</p>
<p>Similarly, when a large object, such as a meteorite or aeroplane, impacts violently at the surface of an ocean, it generates large surface waves, and a family of sound waves that come from a sudden change in pressure known as acoustic-gravity waves. These can travel thousands of kilometres through the water, carrying vital information on the source of the impact, before dissipating. </p>
<p>In <a href="https://www.nature.com/articles/s41598-017-14177-3">our last study</a> we looked at acoustic-gravity waves picked up by hydrophone (underwater microphone) stations in the Indian Ocean, to narrow down where flight MH370 may have impacted the ocean to two points. But now <a href="https://www.nature.com/articles/s41598-018-37626-z">we have found another factor</a> that may prove crucial for pinning down the location of the impact: sea floor elasticity (flexibility). </p>
<p>When acoustic-gravity waves start travelling through the the sea floor their propagation speed boosts to over 3,500 m/s, from the 1,500m/s they would have been travelling at through the water. Previous analysis considered the sea floor to be rigid, which would not allow the radiating waves to move through it. However, if the elasticity of the sea floor is taken into account then the waves will travel at this enhanced speed.</p>
<h2>Rethinking impact</h2>
<p>The acoustic-gravity waves that we analysed for both this and our previous study came from <a href="https://www.ctbto.org/verification-regime/monitoring-technologies-how-they-work/hydroacoustic-monitoring/">two hydroacoustic stations</a> (each of which has three underwater microphones called hydrophones), which were active at the time when MH370 went missing, on March 7-8, 2014. The first, HA01, is off Cape Leeuwin, Western Australia, while the second, known as HA08s, is at Diego Garcia, which is part of the Chagos Archipelago.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/256064/original/file-20190129-108355-1rtdb1g.JPG?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="" src="https://images.theconversation.com/files/256064/original/file-20190129-108355-1rtdb1g.JPG?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/256064/original/file-20190129-108355-1rtdb1g.JPG?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=455&fit=crop&dpr=1 600w, https://images.theconversation.com/files/256064/original/file-20190129-108355-1rtdb1g.JPG?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=455&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/256064/original/file-20190129-108355-1rtdb1g.JPG?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=455&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/256064/original/file-20190129-108355-1rtdb1g.JPG?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=571&fit=crop&dpr=1 754w, https://images.theconversation.com/files/256064/original/file-20190129-108355-1rtdb1g.JPG?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=571&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/256064/original/file-20190129-108355-1rtdb1g.JPG?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=571&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">Map of hydrophone signals recorded on March 7 and 8 2014 between 23:00 and 04:00 UTC, with possible new source locations and two possible MH370 routes.</span>
<span class="attribution"><span class="source">Data SIO, NOAA, U.S. Navy, NGA, GEBCO; © 2018 Basarsoft; US Dept of State Geographer;
© 2018 Google.</span>, <span class="license">Author provided</span></span>
</figcaption>
</figure>
<p><a href="https://www.nature.com/articles/s41598-017-14177-3">Previous studies</a> have <a href="http://www.mh370.gov.my/en/2-uncategorised/69-the-7th-arc">mostly looked at</a> the signals collected by station HA01 between 00:00-02:00 UTC on March 8 2014, as well as signals that related to the last satellite data transmission from MH370 at 00:19 UTC (<a href="http://www.mh370.gov.my/en/2-uncategorised/69-the-7th-arc">known as the 7th arc</a>. However, with our new understanding of acoustic-gravity waves we decided to look at hydroacoustic data from HA01 that was recorded during a wider timeframe – between 23:00 and 04:00 (+1 UTC) on March 7 and 8 – and analysed data from the further away HA08s station too.</p>
<p>Allowing for the effects of sea floor elasticity, the signal locations that we had previously identified using data from HA01 were now different. If the signal travels, say, at twice the speed for a given distance, it should have gone twice the originally calculated distance (without elasticity), so the impact location would be further away relative to the hydrophone station. This is shown on Figure 1 above as signal marks on the purple bearings (the direction of the signals relative to the hydrophones). </p>
<h2>HA08s signals</h2>
<p>Looking at HA08s, the signals were more challenging to analyse. They were distorted by noise which is believed to have been caused by a military exercise in that side of the ocean (depicted as red lines on the map above). However, although the proposed route and point of impact is distant from the 7th arc, we still recommend further studying a number of signals from HA08s. The bearings of some of these signals fall within the area where signals from the military action were picked up, so it is possible that the signals are associated with the military action. But if the signals are related to MH370, this would suggest a new possible impact location in the northern part of the Indian Ocean (as depicted in the top left of the map above).</p>
<p>The locations of signals found using HA08s data do come with high uncertainty but still require further detailed and careful analysis. Unfortunately, on top of the noisy recorded signals, 25 minutes of data from HA08s is missing. The signals we have analysed indicate that the there was a 25-minute shutdown that has gone unexplained by the Comprehensive Nuclear-Test-Ban Treaty Organisation, which is responsible for the hydrophone stations.</p>
<p>In light of this research, we recommended that signals recorded at all times between 23:00 (March 7) and 04:00 (March 8) UTC, at both stations HA01 and HA08s are analysed with no exception. And that this is done independently from other sources (such as satellite data), to minimise inclusion of uncertainties related to them. These recommendations have been communicated to the MH370 Safety Investigation Team in Malaysia, the Australian Transport Safety Bureau, and other relevant authorities with the hope that the search will be resumed to find the missing aircraft.</p><img src="https://counter.theconversation.com/content/110664/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Usama Kadri 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>
Data from hydrophones in the Indian ocean has raised new questions about what happened to MH370.
Usama Kadri, Lecturer of Applied Mathematics, Cardiff University
Licensed as Creative Commons – attribution, no derivatives.
tag:theconversation.com,2011:article/104314
2018-10-07T18:51:46Z
2018-10-07T18:51:46Z
Travelling overseas? What to do if a border agent demands access to your digital device
<figure><img src="https://images.theconversation.com/files/239434/original/file-20181005-52691-12zqgzn.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">New laws enacted in New Zealand give customs agents the right to search your phone.</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/young-business-woman-walking-airport-terminal-1191018481?src=zjpK9WjS-B00bny445xlwQ-1-90">Shutterstock</a></span></figcaption></figure><p><em>New laws enacted in New Zealand this month give border agents the right to demand travellers entering the country hand over passwords for their digital devices. We outline what you should do if it happens to you, in the first part of a <a href="https://theconversation.com/au/topics/tourism-and-technology-60904">series</a> exploring how technology is changing tourism.</em></p>
<hr>
<p>Imagine returning home to Australia or New Zealand after a long-haul flight, exhausted and red-eyed. You’ve just reclaimed your baggage after getting through immigration when you’re stopped by a customs officer who demands you hand over your smartphone and the password. Do you know your rights? </p>
<p>Both Australian and New Zealand customs officers are legally allowed to search not only your personal baggage, but also the contents of your smartphone, tablet or laptop. It doesn’t matter whether you are a citizen or visitor, or whether you’re crossing a border by air, land or sea. </p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/how-to-protect-your-private-data-when-you-travel-to-the-united-states-73909">How to protect your private data when you travel to the United States</a>
</strong>
</em>
</p>
<hr>
<p>New <a href="http://www.legislation.govt.nz/act/public/2018/0004/latest/whole.html">laws</a> that came into effect in New Zealand on October 1 give border agents:</p>
<blockquote>
<p>…the power to make a full search of a stored value instrument (including power to require a user of the instrument to provide access information and other information or assistance that is reasonable and necessary to allow a person to access the instrument).</p>
</blockquote>
<p>Those who don’t comply could face prosecution and NZ$5,000 in fines. Border agents have similar <a href="http://www8.austlii.edu.au/cgi-bin/viewdoc/au/legis/cth/consol_act/ca1901124/s186.html">powers</a> in Australia and elsewhere. In Canada, for example, hindering or obstructing a border guard could cost you up to <a href="https://www.huffingtonpost.ca/2018/08/03/border-guards-phone-search_a_23495655/">C$50,000 or five years in prison</a>. </p>
<h2>A growing trend</h2>
<p>Australia and New Zealand don’t currently publish data on these kinds of searches, but there is a growing trend of device search and seizure at US borders. There was a more than <a href="https://www.nytimes.com/2017/02/14/business/border-enforcement-airport-phones.html">fivefold</a> increase in the number of electronic device inspections between 2015 and 2016 – bringing the total number to 23,000 per year. In the <a href="https://www.cbp.gov/newsroom/national-media-release/cbp-releases-statistics-electronic-device-searches-0">first six months</a> of 2017, the number of searches was already almost 15,000. </p>
<p>In some of these instances, people have been threatened with arrest if they didn’t hand over passwords. Others have been <a href="https://www.cbc.ca/news/alain-philippon-phone-password-case-powers-of-border-agents-and-police-differ-1.2983841">charged</a>. In cases where they did comply, people have <a href="http://www.abc.net.au/news/2018-06-14/airport-searches-of-your-phone-and-computer/9866134">lost sight of their device</a> for a short period, or devices were confiscated and returned days or weeks later.</p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/encrypted-smartphones-secure-your-identity-not-just-your-data-91715">Encrypted smartphones secure your identity, not just your data</a>
</strong>
</em>
</p>
<hr>
<p>On top of device searches, there is also canvassing of social media accounts. In 2016, the United States introduced an <a href="https://www.theguardian.com/world/2016/dec/26/us-customs-social-media-foreign-travelers">additional question</a> on online visa application forms, asking people to divulge <a href="https://www.esta-online.com/blogs/esta-social-media-question">social media usernames</a>. As this form is usually filled out after the flights have been booked, travellers might feel they have no choice but to part with this information rather than risk being denied a visa, despite the question being optional. </p>
<h2>There is little oversight</h2>
<p>Border agents may have a legitimate reason to search an incoming passenger – for instance, if a passenger is suspected of carrying illicit goods, banned items, or agricultural products from abroad. </p>
<p>But searching a smartphone is different from searching luggage. Our smartphones carry our innermost thoughts, intimate pictures, sensitive workplace documents, and private messages. </p>
<p>The practice of searching electronic devices at borders could be compared to police having the <a href="https://www.legislation.gov.au/Details/C2017C00192">right to intercept private communications</a>. But in such cases in Australia, police require a <a href="https://works.bepress.com/kmichael/51/download/">warrant</a> to conduct the intercept. That means there is oversight, and a <a href="https://www.loc.gov/law/help/encrypted-communications/australia.php">mechanism</a> in place to guard against abuse. And the suspected crime must be proportionate to the action taken by law enforcement.</p>
<h2>What to do if it happens to you</h2>
<p>If you’re stopped at a border and asked to hand over your devices and passwords, make sure you have educated yourself in advance about your rights in the country you’re entering. </p>
<p>Find out whether what you are being asked is optional or not. Just because someone in a uniform asks you to do something, it does not necessarily mean you have to comply. If you’re not sure about your rights, ask to speak to a lawyer and don’t say anything that might incriminate you. Keep your cool and don’t argue with the customs officer. </p>
<hr>
<p>
<em>
<strong>
Read more:
<a href="https://theconversation.com/how-secure-is-your-data-when-its-stored-in-the-cloud-90000">How secure is your data when it's stored in the cloud?</a>
</strong>
</em>
</p>
<hr>
<p>You should also be smart about how you manage your data generally. You may wish to switch on two-factor authentication, which requires a password on top of your passcode. And store sensitive information in the cloud on a secure European server while you are travelling, accessing it only on a needs basis. Data protection is taken more seriously in the European Union as a result of the recently enacted <a href="https://eugdpr.org/">General Data Protection Regulation</a>. </p>
<p><a href="https://www.microsoft.com/en-us/servicesagreement">Microsoft</a>, <a href="https://www.apple.com/privacy/manage-your-privacy/">Apple</a> and <a href="https://policies.google.com/terms?hl=en&gl=ZZ">Google</a> all indicate that handing over a password to one of their apps or devices is in breach of their services agreement, privacy management, and safety practices. That doesn’t mean it’s wise to refuse to comply with border force officials, but it does raise questions about the position governments are putting travellers in when they ask for this kind of information.</p><img src="https://counter.theconversation.com/content/104314/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Katina Michael receives funding from the Australian Research Council (ARC). She is affiliated with the Institute of Electrical and Electronics Engineers (IEEE) and the Australian Privacy Foundation (APF).</span></em></p>
Searching a smartphone is different from searching luggage. Our smartphones carry our innermost thoughts, intimate pictures, sensitive workplace documents and private messages.
Katina Michael, Professor, School for the Future of Innovation in Society & School of Computing, Informatics and Decision Systems Engineering, Arizona State University
Licensed as Creative Commons – attribution, no derivatives.
tag:theconversation.com,2011:article/102240
2018-09-04T10:32:52Z
2018-09-04T10:32:52Z
How will Google’s innovation continue beyond its 20th year?
<figure><img src="https://images.theconversation.com/files/234477/original/file-20180831-195322-1423p34.jpg?ixlib=rb-1.1.0&rect=108%2C131%2C3538%2C3306&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">The past and present of Google – what's next?</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/bangkok-thailand-january-24-2017-woman-572139568">Sirirat/Shutterstock.com</a></span></figcaption></figure><p>As millions of people <a href="https://www.w3.org/History.html">came online in the late 1990s</a> they needed help figuring out what each webpage was about, and how to find what they were looking for. Web indexes and search engines sprang up. When Google was founded in September 1998, it had to compete with the <a href="https://pdfs.semanticscholar.org/8395/6915975520e4941610b0cf683ab7bd8521d4.pdf">information retrieval algorithms and techniques</a> – nicknamed “<a href="https://www.searchdecoder.com/ingredients-googles-secret-sauce">secret sauce</a>” – used by <a href="https://gizmodo.com/how-lycos-almost-won-the-search-engine-wars-1719546124">Lycos, Yahoo and other companies</a>. </p>
<p>Technically speaking, Google added two innovations: highly efficient processes for crawling webpages to index their text, and a <a href="https://doi.org/10.1016/S0169-7552(98)00110-X">new way of ranking a page’s relevance</a> based on the number and quality of pages that linked to it. In addition, its interface was refreshingly clean: In an internet then pervaded by pages with lists of lists, <a href="https://www.nngroup.com/articles/search-visible-and-simple/">Google offered a spare alternative</a>, with just a box to type search terms and a “Search” button.</p>
<p>Even more startling was Google’s confidence in its abilities. The company offered a second button, whimsically labeled “<a href="https://web.archive.org/web/20040820044854/http://www.google.com/">I’m feeling lucky</a>,” which would take users directly to the webpage that was the top result – skipping the step of listing possible search results for a user to choose from. It also sought to be a different kind of technology company, early on adopting a straightforward corporate motto: “<a href="https://gizmodo.com/google-removes-nearly-all-mentions-of-dont-be-evil-from-1826153393">Don’t be evil</a>.” Two decades into Google’s history, the power of search is still paramount: <a href="https://searchenginewatch.com/">Entire businesses and professions</a> are built around crafting internet content that will rise to the top of its search results.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/233705/original/file-20180827-75996-2whnf3.png?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="" src="https://images.theconversation.com/files/233705/original/file-20180827-75996-2whnf3.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/233705/original/file-20180827-75996-2whnf3.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=189&fit=crop&dpr=1 600w, https://images.theconversation.com/files/233705/original/file-20180827-75996-2whnf3.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=189&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/233705/original/file-20180827-75996-2whnf3.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=189&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/233705/original/file-20180827-75996-2whnf3.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=238&fit=crop&dpr=1 754w, https://images.theconversation.com/files/233705/original/file-20180827-75996-2whnf3.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=238&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/233705/original/file-20180827-75996-2whnf3.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=238&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">The earliest version of the Google search engine, as stored by the Internet Archive’s Wayback Machine.</span>
<span class="attribution"><a class="source" href="https://web.archive.org/web/19981111183552/http://google.stanford.edu:80/">Screenshot by The Conversation of Archive.org cache of google.stanford.edu</a>, <a class="license" href="http://creativecommons.org/licenses/by-nd/4.0/">CC BY-ND</a></span>
</figcaption>
</figure>
<p>But there are signs of trouble. The company’s role in <a href="https://theconversation.com/when-will-google-defend-democracy-96838">providing misleading information to U.S. voters</a> is under scrutiny. More than 3,100 Google employees signed a <a href="https://www.nytimes.com/2018/04/04/technology/google-letter-ceo-pentagon-project.html">public letter protesting the use of their work in warfare technologies</a> – and about a dozen of them <a href="https://gizmodo.com/google-employees-resign-in-protest-against-pentagon-con-1825729300">resigned in protest</a>. Even more recently, 1,600 Googlers signed a petition to <a href="https://www.digitaltrends.com/web/sundar-pinchai-google-china-censor/">stop their employer from opening a government-restricted search service</a> in China. Additionally, President Donald Trump has questioned whether its <a href="https://theconversation.com/google-news-serves-conservatives-and-liberals-similar-results-but-favors-mainstream-media-102389">rankings for news stories</a> are fair. What might the next 20 years of Google bring? </p>
<h2>Rapid growth</h2>
<p>Google is used to being under scrutiny. In late July 2004 in Sheffield, England, I recall the buzz the company created at the 27th Annual Association of Computing Machinery Special Interest Group in Information Retrieval Research Conference. There were betting pools about when Google would offer its stock for public purchase, and at what price. The Google employees were easy to spot, only using their laptops while sitting with their backs to a wall, so nobody could see what they were reading or typing. </p>
<p>The company <a href="https://www.google.com/about/our-story/">founded by two Stanford graduate students</a> in 1998, which went public on <a href="https://www.investopedia.com/articles/active-trading/081315/if-you-would-have-invested-right-after-googles-ipo.asp">Aug. 19, 2004, at US$85 a share</a>, still gets the <a href="https://www.washingtonpost.com/news/the-switch/wp/2018/02/01/google-parent-alphabet-reports-soaring-ad-revenue-despite-youtube-backlash/">vast majority of its annual revenue</a> from selling search-related advertising. </p>
<p>Yet Google has grown too, in part thanks to a policy giving employees the freedom to work <a href="https://mashable.com/2018/05/11/google-20-percent-rule/">one day a week on side projects</a> that catch their fancy. Now reorganized into an <a href="https://abc.xyz/">umbrella company called Alphabet</a>, the company has expanded into industries as diverse as smartphone operating systems, mapping apps and self-driving vehicles. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/234908/original/file-20180904-45172-1k64m2x.png?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="" src="https://images.theconversation.com/files/234908/original/file-20180904-45172-1k64m2x.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/234908/original/file-20180904-45172-1k64m2x.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=446&fit=crop&dpr=1 600w, https://images.theconversation.com/files/234908/original/file-20180904-45172-1k64m2x.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=446&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/234908/original/file-20180904-45172-1k64m2x.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=446&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/234908/original/file-20180904-45172-1k64m2x.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=560&fit=crop&dpr=1 754w, https://images.theconversation.com/files/234908/original/file-20180904-45172-1k64m2x.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=560&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/234908/original/file-20180904-45172-1k64m2x.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=560&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">Google’s ad revenue from 2001 to 2017 (in billion U.S. dollars).</span>
<span class="attribution"><a class="source" href="https://www.statista.com/statistics/266249/advertising-revenue-of-google/">Statista</a></span>
</figcaption>
</figure>
<h2>Diversification</h2>
<p>Many of the company’s efforts to diversify build on strengths it has developed providing search, such as <a href="https://www.theverge.com/2016/7/21/12246258/google-deepmind-ai-data-center-cooling">cloud computing systems</a> that take advantage of Google engineers’ experience <a href="https://www.wired.com/2012/10/ff-inside-google-data-center/">managing massive data centers</a> and huge amounts of traffic from – and to – sites all around the world.</p>
<p>The company’s massive index of information in many languages is what enabled Google to build a machine-translation system between <a href="https://translate.google.com/intl/en/about/languages/">any of 100 languages</a>. That will help Google remain globally valuable even as <a href="https://www.theverge.com/2018/8/7/17660364/baidu-ceo-google-china">Baidu dominates Chinese-language searches</a>.</p>
<p>Google’s future depends on continuing to create and leverage indexes on features beyond the words on webpages. Combining the ability to identify a user performing the search with its knowledge of that person’s search history and their current location, Google can already provide <a href="https://www.pcworld.com/article/3299042/privacy/google-privacy-checkup-faq.html">finely tuned personalized results</a>. A new company effort is already planning to use <a href="https://slate.com/technology/2018/08/google-coach-is-google-going-to-help-you-lose-weight-and-eat-better.html">health devices people wear, implant or carry on their bodies</a> to provide useful nutrition and fitness tips.</p>
<p>Google is no doubt planning to add to its special sauce indexes of social media posts, data from <a href="https://techcrunch.com/2017/05/11/google-funded-super-sensor-project-brings-iot-powers-to-dumb-appliances/">sensors in the environment</a> – including cameras, microphones and all sorts of connected “internet of things” devices.</p>
<h2>Future challenges</h2>
<p>Google is already applying its expertise to its line of <a href="https://slate.com/technology/2018/08/amazon-echo-is-losing-smart-speaker-market-share-to-google-home-heres-why.html">smart speakers and personal assistants</a>, offering its well-regarded search results through voice recognition and spoken responses. One day typing text onto a screen may seem as quaint as rotary phones.</p>
<p>A next category of features might be called anticipatory search, providing information or suggesting action without a user even specifying a query. For instance, some cars already go beyond alerting the driver to low fuel levels, locating and <a href="https://www.lexusofroseville.com/FeaturesTechnology_D?p=2019_NX">providing directions to nearby gas stations</a>. One day a personal fitness tracker might note that a user’s resting heart rate is 15 percent higher this week than the average over the past six months. From there, it might offer up research or doctors’ advice about cardiovascular health. </p>
<p><a href="https://developers.google.com/recaptcha/"><img src="https://cdn.theconversation.com/static_files/files/254/newCaptchaAnchor.gif?1535735401"></a></p>
<p>Google may even ramp up its efforts to distinguish people from machines – such as “<a href="https://www.google.com/recaptcha/intro/v3beta.html">captcha</a>” challenges and <a href="https://support.google.com/accounts/answer/1066447?co=GENIE.Platform%3DAndroid&hl=en">multi-factor</a> <a href="https://venturebeat.com/2018/08/30/hands-on-with-the-titan-security-key-a-50-fido-fob-that-secures-your-online-accounts/">authentication processes</a>. From there, it may work to eliminate the increasing efforts from both humans and computers – such as <a href="https://theconversation.com/how-the-russian-government-used-disinformation-and-cyber-warfare-in-2016-election-an-ethical-hacker-explains-99989">Russian government agents</a> and <a href="https://theconversation.com/how-twitter-bots-affected-the-us-presidential-campaign-68406">Twitter bots</a> – to secretly influence search results for malicious purposes.</p>
<p>These features may sound exciting and useful, but they also carry <a href="https://theconversation.com/facebooks-social-responsibility-should-include-privacy-protection-94549">important ethical concerns</a>, about who can access people’s personal data, and for what purposes. It will be interesting to see whether the concerns Google employees are currently expressing about political uses of their work will extend to personal privacy, and whether – and how – any objections might influence searches in the future.</p><img src="https://counter.theconversation.com/content/102240/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Gary Marchionini owns a small amount of Google stock (2 shares I bought for my grandchildren) and of course through various funds in my retirement accounts.
I have had grants from NSF to study information retrieval research over the past 25 years and my students and I have had research grants from Google (most recent was about 5 years ago).</span></em></p>
As Google turns 20, a look at how the company has grown – and what the next two decades might bring for the company.
Gary Marchionini, Professor of Information and Library Science, University of North Carolina at Chapel Hill
Licensed as Creative Commons – attribution, no derivatives.
tag:theconversation.com,2011:article/92791
2018-03-23T18:56:55Z
2018-03-23T18:56:55Z
‘Big Tech’ isn’t one big monopoly – it’s 5 companies all in different businesses
<figure><img src="https://images.theconversation.com/files/211745/original/file-20180323-54898-1dnsu0o.png?ixlib=rb-1.1.0&rect=1029%2C0%2C1844%2C1255&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">It may seem convenient to think of technology companies as similar, but they're really not.</span> <span class="attribution"><span class="source">The Conversation</span>, <a class="license" href="http://creativecommons.org/licenses/by-nc/4.0/">CC BY-NC</a></span></figcaption></figure><p>Public <a href="https://www.theatlantic.com/technology/archive/2018/03/facebook-cambridge-analytica/555866/">concern</a> about Facebook’s power in society – and in politics – has skyrocketed in the wake of <a href="https://www.theguardian.com/news/2018/mar/17/cambridge-analytica-facebook-influence-us-election">revelations</a> that users’ data was analyzed by a U.K.-based marketing firm and used to construct highly targeted political propaganda in advance of the 2016 U.S. presidential election. Other technology giants have also sparked concern: <a href="https://www.ft.com/content/9554a8bc-5b12-11e7-b553-e2df1b0c3220">Google</a>, <a href="https://www.cnbc.com/2017/05/08/apple-google-european-commission-spotify.html">Apple</a>, <a href="https://www.cnbc.com/2018/02/05/tech-investor-warns-amazon-against-abusing-its-power-to-influence-users.html">Amazon</a> and <a href="http://knowledge.wharton.upenn.edu/article/microsoft-a-case-of-justice-or-abuse-of-power/">Microsoft</a> have all faced objections from users, the public and even government agencies. </p>
<p>Because all of these companies provide services relating to computers, there is a tendency to lump them together, calling them “<a href="https://www.nytimes.com/2015/09/20/opinion/is-big-tech-too-powerful-ask-google.html">Big Tech</a>” or the “<a href="https://www.nytimes.com/2017/10/11/technology/the-frightful-five-want-to-rule-entertainment-they-are-hitting-limits.html">Frightful Five</a>” or even “<a href="https://qz.com/303947/us-cultural-imperialism-has-a-new-name-gafa/">GAFA</a>” – the acronym for the first four of them, leaving Microsoft out. Conceiving of “<a href="http://www.slate.com/articles/technology/technology/2017/11/how_silicon_valley_became_big_tech.html">big tech</a>” as a single industry makes the threat and influence overwhelming. </p>
<p>In the U.S., when an industry gets so large it exerts political pressure on society, people often label the industry as a whole, like “<a href="https://content.time.com/time/magazine/article/0,9171,920328,00.html">Big Oil</a>,” “<a href="https://www.huffingtonpost.com/entry/big-tobacco-is-still-in-the-business-of-deceiving-americans_us_5a202d96e4b0392a4ebbf5f3">Big Tobacco</a>” or “<a href="https://www.thedailybeast.com/big-pharma-is-americas-new-mafia">Big Pharma</a>.” The so-called big tech companies certainly are big: In 2017, they were the top five <a href="https://247wallst.com/investing/2017/10/29/market-cap-of-5-largest-us-companies-up-36-in-most-recent-year/">most valuable public companies</a> in the U.S. But, as a <a href="https://mitpress.mit.edu/books/we-now-disrupt-broadcast">scholar of the media marketplace</a> that many of these firms are beginning to explore, I know that lumping them together hides the fact they’re very separate and distinct – not just as companies, but in terms of their business models and practices.</p>
<p>Understanding these companies in their proper business contexts makes it easier to understand their power in the marketplace and society at large. It also suggests ways to assess, regulate and manage that power to protect competition and <a href="https://theconversation.com/facebook-is-killing-democracy-with-its-personality-profiling-data-93611">even democracy itself</a>. </p>
<h2>Google: Advertising revenue from searches</h2>
<p>Google and Facebook are most frequently discussed together, likely because of their domination of internet advertising. Together, the two companies <a href="https://www.emarketer.com/Article/Google-Facebook-Tighten-Grip-on-US-Digital-Ad-Market/1016494">collected 63 percent</a> of U.S. digital advertising dollars in 2017. Both companies earn most of their revenue from advertising: 97 percent for <a href="https://www.visualcapitalist.com/chart-5-tech-giants-make-billions/">Facebook</a> and 88 percent for <a href="https://www.investopedia.com/articles/investing/020515/business-google.asp">Google’s</a> parent company Alphabet in 2016. But what they offer to advertisers and what users want from them are very different.</p>
<p>Google’s value proposition is helping users find things. Many – even most – of the <a href="http://www.internetlivestats.com/google-search-statistics/">3.5 billion</a> searches Google performs each day aren’t monetized at all. Google only gets paid if a searcher clicks on a paid link; the top three results are often labeled as “Ads,” in addition to several on the right side of a computer user’s search results screen. </p>
<p>Advertisers like Google because they only pay if their <a href="https://adwords.google.com/home/pricing/">ads are clicked</a>. That is a far better deal than what is offered in traditional media advertising, where payment is for how many people are shown an ad, rather than customers’ responses. In addition, Google’s position as a <a href="https://www.thinkwithgoogle.com/consumer-insights/mobile-search-consumer-behavior-data/">leading place</a> where people look for information on products and services means an ad reaches a consumer exactly at the moment they’re looking for a product. This timing is more valuable than just showing ads to people in general – so much so that advertisers paid Google <a href="https://www.statista.com/statistics/266249/advertising-revenue-of-google/">US$79.38 billion</a> in 2016.</p>
<h2>Facebook: Ad revenue from attention-grabbing content</h2>
<p>Facebook operates more like a traditional ad-supported media company. It provides interesting content that attracts an audience, and sells their attention to advertisers – just as television, radio and print have done for decades. The key difference between Facebook and these legacy media businesses is where the content comes from: Rather than Facebook paying to create the material that draws users, the users add it themselves for free, posting personal messages and shared links. </p>
<p>Like traditional media, Facebook charges advertisers based on how many people see a message, not on how many take action by clicking. The value Facebook offers over traditional advertising is its ability to <a href="https://www.propublica.org/article/facebook-enabled-advertisers-to-reach-jew-haters">target very particular groups</a> with a <a href="https://www.marketplace.org/2010/11/26/tech/marketers-you">customized advertising message</a>. This is precisely the type of <a href="https://www.cnet.com/news/facebook-cambridge-analytica-data-mining-and-trump-what-you-need-to-know/">targeting</a> that happened during the 2016 U.S. presidential election, which generated widespread public criticism.</p>
<h2>Apple: Selling electronic hardware</h2>
<p>In contrast to the advertising businesses of Google and Facebook, Apple remains a hardware technology company, deriving <a href="https://www.visualcapitalist.com/chart-5-tech-giants-make-billions/">84 percent</a> of its 2016 revenue from the iPhone, iPad and Mac computers. The profits on those sales let Apple use <a href="https://www.fastcompany.com/1784824/great-tech-war-2012">very different strategies</a> than the non-hardware companies with which it is often compared. The profit margins on each device are so substantial it <a href="https://www.fastcompany.com/1784824/great-tech-war-2012">doesn’t have to dominate</a> the hardware market the way Google and Facebook control online advertising. Despite the seeming ubiquity of iPhones in some social circles, iPhones <a href="https://www.statista.com/statistics/216459/global-market-share-of-apple-iphone/">rarely top 20 percent</a> of worldwide phone sales, and account for <a href="https://9to5mac.com/2017/08/09/us-iphone-sales-ios-market-share-kantar/">about 30 percent</a> of U.S. sales.</p>
<p>Apple has other elements to its business, too – such as its iTunes music distribution business. But it’s important to keep the relative scale of those elements in mind. Mostly, they are <a href="https://www.thecontenttrap.com/">complementary businesses</a> that Apple uses strategically in support of its primary focus as a hardware company. Taken together, iTunes, its App Stores, iBooks Store, Apple Music, Apple Care, Apple Pay and other even more ancillary sales added up to <a href="http://investor.apple.com/secfiling.cfm?filingID=1628280-16-20309&CIK=320193#A201610-K9242016_HTM_SE78948B641FF55EDB70F7F75DDCB7673">just 11 percent</a> of the company’s revenue in 2016. Even the company’s plan to spend <a href="http://variety.com/2017/digital/news/apple-1-billion-original-tv-shows-movies-budget-1202529421/">$1 billion on original video</a> is hard to understand, except as a support to branding and marketing efforts that boost its hardware sales.</p>
<h2>Microsoft and Amazon: Mixed retail, computing and media</h2>
<p>Much like Apple, Microsoft blends many revenue streams: It sells Surface computers, Azure cloud services, software (like the Microsoft Office Suite), gaming consoles and search engine advertising. The company once stood alone as a <a href="http://eprints.lse.ac.uk/47043/1/CentrePiece_12_1.pdf">poster child</a> for massive technology corporations. Lately, it may draw less attention because competitors like Google’s G Suite have challenged its market share. Also, Microsoft has not aggressively entered social media, a sector now under great scrutiny.</p>
<p>Finally, Amazon also operates in many different business sectors. Primarily, it is a goods retailer: That’s where <a href="https://www.visualcapitalist.com/chart-5-tech-giants-make-billions/">70 percent</a> of its annual revenue came from in 2016. Its Amazon Web Services content hosting and cloud computing business contributed 9 percent, and Amazon’s media businesses provided roughly 18 percent of the company’s $136 billion of annual revenue. That $24 billion of media revenue is nearly three times that of Netflix, but still not Amazon’s core business.</p>
<h2>Regulate markets and behavior, not ‘tech’</h2>
<p>It’s not that these companies are so different as to be unrelated or incomparable to each other. They all involve – to varying degrees – computers and services built on internet connection that provide services to customers in ways that never existed before. All five gather data on their users and analyze behavior using algorithms to create personal experiences in ways that are new and have been challenging for companies with long histories in sectors such as media, transportation or retail to match.</p>
<p>But despite simple perception of them all as “<a href="https://www.inc.com/magazine/201605/marli-guzzetta/tech-company-definition.html">tech</a>” companies, their core revenue sources are clearly different. And those distinctions suggest ways people can understand and respond to anxieties about their growing economic and cultural influence.</p>
<p>In fact, what is most concerning is the extent to which these companies aren’t in the same businesses: They’re not competing with each other, or really anyone else.</p>
<p>In prior eras, Americans learned that major industries they first viewed as innovators and economic saviors were more complicated and less magnanimous than initially believed. So now today, big tech isn’t unlike everything that came before. In fact, big tech <a href="https://al3x.net/2012/05/08/what-is-and-is-not-a-technology-company.html">isn’t really a thing</a> at all. Assessing these companies based on what they do, rather than mythologizing them, is the first step forward.</p>
<p>
<section class="inline-content">
<img src="https://images.theconversation.com/files/248895/original/file-20181204-133100-t34yqm.png?w=128&h=128">
<div>
<header>Amanda Lotz is the author of:</header>
<p><a href="https://mitpress.mit.edu/books/we-now-disrupt-broadcast">We Now Disrupt This Broadcast:
How Cable Transformed Television and the Internet Revolutionized It All</a></p>
<footer>MIT Press provides funding as a member of The Conversation US.</footer>
</div>
</section>
</p><img src="https://counter.theconversation.com/content/92791/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Amanda Lotz 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>
When thinking about regulating them, it’s useful to know Facebook, Amazon, Google, Apple and Microsoft have some similarities. But generally they’re not competing with each other – or anyone else.
Amanda Lotz, Fellow, Peabody Media Center; Professor of Media Studies, University of Michigan
Licensed as Creative Commons – attribution, no derivatives.
tag:theconversation.com,2011:article/86002
2017-10-24T09:03:26Z
2017-10-24T09:03:26Z
New tool to track underwater acoustic waves could find MH370
<figure><img src="https://images.theconversation.com/files/191482/original/file-20171023-1738-1a9ik4p.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-photo/underwater-wave-733264549?src=90_EqEFMrsHt2B4oJ9xi2w-1-7">Melissa Burovac/Shutterstock</a></span></figcaption></figure><p>Looking at the ocean, a lake, or even a pond, you may wonder what happens to the waves you see when they “disappear”. These surface waves tend to become smaller and smaller until you can’t see them anymore. But they keep travelling through the water at a lower depth. These “<a href="http://www.sciencedirect.com/science/article/pii/S0997754614001253">acoustic-gravity waves</a>” can travel for thousands of kilometres undisturbed, and even cross an entire ocean. </p>
<p>These compression waves are generated by a sudden change in the water pressure. They can be caused by anything from submarines, earthquakes and landslides, to falling meteorites or other objects impacting the sea surface. And although they are “acoustic” waves, they are below the range of human hearing – the only way to pick up and record them is using hydrophones, special microphones that work underwater. </p>
<p>The <a href="https://www.ctbto.org/">Comprehensive Nuclear-Test-Ban Treaty Organisation</a> (CTBTO) has hydrophone stations dotted in oceans across the world. They are used by the organisation to detect shock waves that may be a consequence of an underwater nuclear test – but we have found <a href="http://www.nature.com/articles/s41598-017-14177-3">a way to use these signals</a>, to find where and when acoustic waves are originally generated.</p>
<h2>Searching with sound</h2>
<p>Some of the organisation’s stations, such as HA01, located off Cape Leeuwin in south-west Australia, have three hydrophones. This configuration lets us calculate the direction of the waves quite accurately because the incoming waves hit the hydrophones in a particular order, similar to how soundwaves hit human ears. But unlike sounds processed by our brain, hydrophones alone cannot easily tell how far away the event was generated, or what generated it.</p>
<p>To do this, we used mathematical tools which consider the way acoustic-gravity waves behave. As these waves travel through the water, <a href="http://physics.usask.ca/%7Ehirose/ep225/animation/dispersion/anim-dispersion.html">they disperse</a>. This means that groups of waves created by a source start off being close together, but tend to become more spread apart as they travel further – this is because lower frequency soundwaves are a bit slower than those at higher frequency. By looking at how frequencies disperse, we can estimate how far the wave has travelled, and this can give us an estimate of where they originated from.</p>
<p>Our study was initially motivated by a desire to gain more knowledge about <a href="http://avherald.com/h?article=4710c69b">the incident involving missing flight MH370</a>. To confirm that our idea worked we targeted two 5.1 magnitude earthquakes, which had already been localised by seismometers, and tried to find their location with our method. The accuracy was quite good (with errors of around 100-150km), considering that the signals travelled for 2,000km in one case and 5,000km in the other, and that the hydrophones picked up other noises due to surface wind, boats, and other underwater sources.</p>
<figure class="align-center ">
<img alt="" src="https://images.theconversation.com/files/191025/original/file-20171019-1075-1n0w7ee.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/191025/original/file-20171019-1075-1n0w7ee.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=629&fit=crop&dpr=1 600w, https://images.theconversation.com/files/191025/original/file-20171019-1075-1n0w7ee.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=629&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/191025/original/file-20171019-1075-1n0w7ee.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=629&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/191025/original/file-20171019-1075-1n0w7ee.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=790&fit=crop&dpr=1 754w, https://images.theconversation.com/files/191025/original/file-20171019-1075-1n0w7ee.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=790&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/191025/original/file-20171019-1075-1n0w7ee.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=790&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">Location of earthquakes, and of signal thought to be from an object impacting the surface.</span>
<span class="attribution"><span class="source">Davide Crivelli/Usama Kadri</span>, <span class="license">Author provided</span></span>
</figcaption>
</figure>
<p>In our search we also found a very interesting signal coming from an area near the Antarctic circle. This signal looks surprisingly similar to one which we obtained by dropping a heavy sphere in a large tank, 40 metres deep, during a set of experiments. We think the ocean signal could have been caused by a meteorite, but have yet to confirm this with NASA.</p>
<figure class="align-center ">
<img alt="" src="https://images.theconversation.com/files/191027/original/file-20171019-1075-1myowh4.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/191027/original/file-20171019-1075-1myowh4.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=551&fit=crop&dpr=1 600w, https://images.theconversation.com/files/191027/original/file-20171019-1075-1myowh4.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=551&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/191027/original/file-20171019-1075-1myowh4.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=551&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/191027/original/file-20171019-1075-1myowh4.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=692&fit=crop&dpr=1 754w, https://images.theconversation.com/files/191027/original/file-20171019-1075-1myowh4.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=692&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/191027/original/file-20171019-1075-1myowh4.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=692&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">Top figure shows the signature recorded in Canadian experiments; the bottom three signals were captured on the HA01 hydrophones.</span>
<span class="attribution"><span class="source">Davide Crivelli/Usama Kadri</span>, <span class="license">Author provided</span></span>
</figcaption>
</figure>
<h2>MH370</h2>
<p>Since confirming that the technique worked, we have used advanced automated methods to find signals buried inside the hours of data recorded by the hydrophone station off Cape Leeuwin before the time flight MH370 was believed to have <a href="https://theconversation.com/a-year-after-mh370-vanished-we-have-only-theories-but-heres-why-the-search-must-go-on-38348">run out of fuel</a>. We were able to find and localise two very faint signals – one ten minutes after the last satellite communication with the plane – but far from the probable location arc, and another almost one hour later, closer to the last area where the plane last communicated with a satellite. </p>
<figure class="align-center ">
<img alt="" src="https://images.theconversation.com/files/191028/original/file-20171019-1072-er7nmx.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/191028/original/file-20171019-1072-er7nmx.png?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=584&fit=crop&dpr=1 600w, https://images.theconversation.com/files/191028/original/file-20171019-1072-er7nmx.png?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=584&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/191028/original/file-20171019-1072-er7nmx.png?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=584&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/191028/original/file-20171019-1072-er7nmx.png?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=734&fit=crop&dpr=1 754w, https://images.theconversation.com/files/191028/original/file-20171019-1072-er7nmx.png?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=734&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/191028/original/file-20171019-1072-er7nmx.png?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=734&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">The two events (E1 and E2) captured on March 8 2014, between 00:00 UTC and 02:00 UTC.</span>
<span class="attribution"><span class="source">Davide Crivelli/Usama Kadri</span>, <span class="license">Author provided</span></span>
</figcaption>
</figure>
<p>Though we have located two points around the time of MH370’s disappearance, we cannot say with any real certainty that these have any association with the aircraft. Just like in a busy restaurant, it gets more and more difficult to pick up individual voices as the noise in the room gets louder. What we do know is that the hydrophones picked up remarkably weak signals at these locations and that the signals, according to our calculations, accounted for some sort of source in the Indian Ocean.</p>
<p>All of this information has been passed onto the Australian Transport Safety Bureau – the government body which was <a href="https://www.atsb.gov.au/mh370-pages/the-search/about-the-search/">leading the search for MH370</a> until it was suspended on January 17 2017. We anticipate that both now, and in the future, this new source of information could be <a href="https://theconversation.com/new-research-points-to-the-crash-site-of-missing-plane-mh370-63019">used in conjunction</a> with a whole <a href="https://www.atsb.gov.au/mh370/">host of other data</a> that is at the disposal of the authorities in the search for missing objects at sea.</p><img src="https://counter.theconversation.com/content/86002/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>The authors do not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.</span></em></p>
A new method has been developed to find objects that land at sea using underwater sounds.
Davide Crivelli, Lecturer in Mechanical Engineering, Cardiff University
Usama Kadri, Lecturer of Applied Mathematics, Cardiff University
Licensed as Creative Commons – attribution, no derivatives.
tag:theconversation.com,2011:article/76688
2017-05-05T18:03:54Z
2017-05-05T18:03:54Z
Fake news, echo chambers and filter bubbles: Underresearched and overhyped
<figure><img src="https://images.theconversation.com/files/167927/original/file-20170504-4929-1sx8gvi.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">Don't panic: An international survey finds concerns about fake news are overblown.</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-vector/businessman-working-on-laptop-pop-art-356914784">studiostoks/shutterstock.com</a></span></figcaption></figure><p>In the early years of the internet, it was revolutionary to have <a href="https://www.theguardian.com/technology/2014/aug/24/internet-lost-its-way-tim-berners-lee-world-wide-web">a world of information just a click away</a> from anyone, anywhere, anytime. Many hoped this <a href="http://www.hup.harvard.edu/catalog.php?isbn=9780674872332">inherently democratic technology</a> could lead to better-informed citizens more easily <a href="http://www.nature.com/news/society-build-digital-democracy-1.18690">participating in debate, elections and public discourse</a>.</p>
<p>Today, though, many observers are <a href="https://theconversation.com/is-googles-eagerness-to-answer-questions-promoting-more-falsehood-online-70894">concerned that search algorithms</a> and <a href="https://www.theguardian.com/technology/2016/dec/04/google-democracy-truth-internet-search-facebook">social media are undermining</a> the <a href="https://www.theguardian.com/technology/2016/dec/16/google-autocomplete-rightwing-bias-algorithm-political-propaganda">quality of online information</a> people see. They worry that bad information may be <a href="https://www.researchgate.net/publication/220427075_Why_the_Internet_is_Bad_for_Democracy">weakening democracy in the digital age</a>.</p>
<p>The problems include online services <a href="https://theconversation.com/how-can-we-learn-to-reject-fake-news-in-the-digital-world-69706">conveying fake news</a>, splitting users into “<a href="http://www.penguinrandomhouse.com/books/309214/the-filter-bubble-by-eli-pariser/9780143121237/">filter bubbles</a>” of <a href="https://www.wired.com/2016/11/filter-bubble-destroying-democracy/">like-minded people</a> and <a href="http://press.princeton.edu/chapters/s8468.html">enabling users to unwittingly</a> <a href="https://theconversation.com/why-do-we-fall-for-fake-news-69829">lock themselves up</a> in <a href="https://www.psychologytoday.com/blog/psych-unseen/201611/fake-news-echo-chambers-filter-bubbles-survival-guide">virtual echo chambers</a> that <a href="http://www.newstatesman.com/helen-lewis/2015/07/echo-chamber-social-media-luring-left-cosy-delusion-and-dangerous-insularity">reinforce their own biases</a>. </p>
<p>These concerns are much discussed, but have not yet been thoroughly studied. <a href="https://doi.org/10.7717/peerj-cs.38">What research does exist</a> has typically been limited to a single platform, such Twitter or Facebook. Our <a href="http://ssrn.com/abstract=2960697">study of search and politics in seven nations</a> – which surveyed the United States, Britain, France, Germany, Italy, Poland and Spain in January 2017 – found these concerns to be overstated, if not wrong. In fact, many internet users trust search to help them find the best information, check other sources and discover new information in ways that can burst filter bubbles and open echo chambers. </p>
<h2>Surveying internet users</h2>
<p>We sought to learn directly from people about how they used search engines, social media and other sources of information about politics. Through funding from Google, we conducted an <a href="http://ssrn.com/abstract=2960697">online survey of more than 14,000 internet users in seven nations</a>. </p>
<p>We found that the fears surrounding search algorithms and social media are not irrelevant – <a href="https://www.sciencenews.org/blog/science-public/youve-probably-been-tricked-fake-news-and-dont-know-it">there are problems for some users some of the time</a>. However, they are exaggerated, creating unwarranted fears that could lead to inappropriate responses by users, regulators and policymakers. </p>
<h2>The importance of searching</h2>
<p>The survey findings demonstrate the importance of search results over other ways to get information. When people are looking for information, they very often search the internet. Nearly two-thirds of users across our seven nations said they use a search engine to look for news online at least once a day. They view search results as equally accurate and reliable as other key sources, like television news.</p>
<p><iframe id="pfmZV" class="tc-infographic-datawrapper" src="https://datawrapper.dwcdn.net/pfmZV/1/" height="400px" width="100%" style="border: none" frameborder="0"></iframe></p>
<p>In line with that general finding, a search engine is the first place internet users go online for information about politics. Moreover, those internet users who are very interested in politics, and who participate in political activities online, are the most likely to use a search engine like Bing or Google to find information online about politics.</p>
<p>But crucially, those same users engaged in search are also very likely to get information about politics on other media, exposing themselves to diverse sources of information, which makes them more likely to encounter diverse viewpoints. Further, we found that people who are interested and involved in politics online are more likely to double-check questionable information they find on the internet and social media, including by searching online for additional sources in ways that will pop filter bubbles and break out of echo chambers.</p>
<h2>Internet-savvy or not?</h2>
<p>It’s not just politically interested people who have these helpful search habits: People who use the internet more often and have more practice searching online do so as well.</p>
<p>That leaves the least politically interested people and the least skilled internet users as most susceptible to fake news, filter bubbles and echo chambers online. These individuals could <a href="http://dx.doi.org/10.1007/978-3-319-40548-3_74">benefit from support</a> and <a href="https://theconversation.com/the-challenge-facing-libraries-in-an-era-of-fake-news-70828">training in digital literacy</a>.</p>
<p>However, for most people, internet searches are critical for checking the reliability and validity of information they come across, whether online, on social media, on traditional media or in everyday conversation. Our research shows that these internet users find search engines useful for checking facts, discovering new information, understanding others’ views on issues, exploring their own views and deciding how to vote.</p>
<h2>International variations</h2>
<p>We found that people in different countries do vary in how much they trust and rely on the internet and searches for information. For example, internet users in Germany, and to a lesser extent those in France and the United Kingdom, are more trusting in TV and radio news, and more skeptical of searches and online information. Internet users in Germany rate the reliability of search engines lower than those in all the other nations, with 44 percent saying search engines are reliable, compared with 50 to 57 percent across the other six countries.</p>
<p><iframe id="nQXkq" class="tc-infographic-datawrapper" src="https://datawrapper.dwcdn.net/nQXkq/2/" height="400px" width="100%" style="border: none" frameborder="0"></iframe></p>
<p>In Poland, Italy and Spain, people trust traditional broadcast media less and are more reliant on, and trusting of, internet and searching. Americans are in the middle; there were greater differences within European countries than between Europe as a whole and the U.S. American internet users were so much more likely to consult multiple sources of information that we called them “media omnivores.”</p>
<p>Internet users generally rely on a diverse array of sources for political information. And they display a healthy skepticism, leading them to question information and check facts. <a href="https://www.theguardian.com/technology/2016/nov/29/facebook-fake-news-problem-experts-pitch-ideas-algorithms">Regulating the internet</a>, as some have proposed, could undermine existing trust and introduce new questions about accuracy and bias in search results.</p>
<p>But panic over fake news, echo chambers and filter bubbles is exaggerated, and not supported by the evidence from users across seven countries.</p><img src="https://counter.theconversation.com/content/76688/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>William H. Dutton received funding from Google through Michigan State University to conduct the survey on which these findings are based. However, Google did not design the study, questionnaire, analysis, or findings. The opinions are those of the author and not any organization that supported this research. </span></em></p>
Concerns over filter bubbles and fake news are often based on anecdotal evidence. There is relatively little systematic research on the topic; a new survey finds widespread fears are unwarranted.
William H. Dutton, Professor of Media and Information Policy, Michigan State University
Licensed as Creative Commons – attribution, no derivatives.
tag:theconversation.com,2011:article/75033
2017-04-03T14:47:50Z
2017-04-03T14:47:50Z
Google’s YouTube battle with big brands could shake up how content is shared
<figure><img src="https://images.theconversation.com/files/163640/original/image-20170403-21938-1c52o4k.jpg?ixlib=rb-1.1.0&rect=3%2C0%2C1000%2C613&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><a class="source" href="https://www.flickr.com/photos/burningmax/2096518838/in/photolist-4cgd1w-9CYEG6-68GHaL-dDuW4C-9shrxM-uBUKwq-ohAYe7-4nWKDi-8s6j1B-dX4wFM-7cymcP-nnEpB-sDRDww-dgZu6y-adaghs-7z3Z8i-deN843-p3Zh9E-7cAKFm-f2qxd-a6xLt-aNtKKt-7mmtWg-oXAjCs-2EyvES-wYXg9-emp7wE-wYXbh-2bCsZi-dgZu4K-53n3x4-7vsL1G-8tsBYc-cYbxaj-wYWZX-6gNoYT-9JqKhB-8s9nsd-7cyJBL-9VtUmi-7cuCkc-e1yuBz-7cy19t-dX4sAR-bF4WiT-fko77Y-7cA4tu-7cA55C-5rSZk-qAYCK1">burningmax/Flickr</a>, <a class="license" href="http://creativecommons.org/licenses/by/4.0/">CC BY</a></span></figcaption></figure><p>Leading brands are boycotting advertising services on Google’s YouTube after their ads were placed alongside content they deemed inappropriate and – even worse – were charged for the privilege. The long list includes household names such as Starbucks, Pepsi, Walmart, Tesco, the BBC and <a href="https://www.theguardian.com/technology/2017/mar/19/google-braces-for-questions-as-more-big-name-firms-pull-adverts?CMP=share_btn_tw">the UK government</a>. Some have even suggested that they may have inadvertently sponsored extremist and hateful groups to the tune of more than <a href="https://www.theguardian.com/technology/2017/mar/17/extremists-ads-uk-brands-google-wagdi-ghoneim?CMP=share_btn_tw">£250,000</a>.</p>
<p>To understand how, it’s helpful to recall how Google services such as YouTube work. When we watch a video clip we may see ads and we often skip them. But for Google the ad is the important thing – <a href="https://www.theguardian.com/technology/2015/aug/10/google-alphabet-parent-company">something like 90% of its revenue</a> comes from selling ads. When an ad is placed around a YouTube video, Google charges the advertiser and shares some of the revenue with the user who posted the video. Vloggers such as <a href="http://www.cosmopolitan.com/uk/entertainment/news/a41923/beauty-vlogger-zoella-earns-thousands-a-month/">Zoella</a> make a pretty penny from this. Google’s targeting technologies can allow advertisers to place their ads around specific types of videos or present them to specific groups of viewers.</p>
<p>So what’s the problem? Surely it is acceptable for brands to choose which content they sponsor? Isn’t this what happens on traditional media? Given that brands can target their ads already perhaps something else is going on.</p>
<h2>Over the firewall</h2>
<p>Brands certainly have the right to demand which content their ads appear alongside. On traditional media, a brand might not want its ads placed alongside a report about a terror campaign, for example. But it is not considered acceptable for them to say that the newspaper shouldn’t report such events at all. To get around this, traditional media companies separate news and information operations from advertising operations.</p>
<p>The problem with services like Google and other social media platforms is that such organisational firewalls <a href="http://triple-c.at/index.php/tripleC/article/viewArticle/255">don’t exist</a>. As a result, there are potential hidden biases that can come about when advertisers insist that content matches their “<a href="https://www.theguardian.com/technology/2017/mar/24/walmart-starbucks-pepsi-pull-ads-google-youtube?CMP=share_btn_tw">company values</a>”. Ironically enough, this problem is spelled out pretty well <a href="http://infolab.stanford.edu/%7Ebackrub/google.html">in a 1998 paper</a> by Google’s founders <a href="https://en.wikipedia.org/wiki/Sergey_Brin">Sergey Brin</a> and <a href="https://en.wikipedia.org/wiki/Larry_Page">Larry Page</a>. </p>
<p>They wrote that the “goals of the advertising business model do not always correspond to providing quality search to users”. The example they cited from a prototype version of Google saw a search for mobile phones pull up a study which went into detail about the risks of speaking on a phone while driving. They concluded that a search engine company which was taking money for mobile phone ads would struggle to justify that kind of search result to paying advertisers. Brin and Page concluded that:</p>
<blockquote>
<p>Advertising funded search engines will be inherently biased towards the advertisers and away from the needs of the consumers.</p>
</blockquote>
<p>That belief underpins why Google search separates out advertising sponsored links from organic search results. On YouTube, however, responsibility for the content of videos and the placement of ads is delegated onto viewers, users and brands through its <a href="https://www.youtube.com/static?gl=GB&template=terms">terms of service</a> and <a href="https://support.google.com/youtube/answer/6162278?hl=en-GB">advertising policies</a>. </p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/BNHR6IQJGZs?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
</figure>
<p>Indeed, from my analysis of 45 published cases by the UK’s <a href="http://www.asa.org.uk">Advertising Standards Authority</a> that involved YouTube, there were 44 in which YouTube responded that users, viewers and brands were responsible for the ads people saw on the service – not them. In just a single case, did YouTube accept that an ad should be taken down following a complaint to the ASA.</p>
<p>So, <a href="http://adage.com/article/digital/wins-loses-youtube-s-falling-brands/308450/?utm_campaign=SocialFlow&utm_source=Twitter&utm_medium=Social">one way</a> of looking at the current boycott is that’s it’s a case of brands pushing responsibility for content and placement back onto Google/YouTube who, up to now, have been very successful in passing this responsibility onto brands and user. </p>
<h2>Weeding</h2>
<p>There is another way of looking at it, and that is as an indirect way for brands to claim more power over the type of content that people can share online. If Google has to “weed out extremist views” (as politicians such as <a href="https://www.theguardian.com/technology/2017/mar/20/google-ads-extremist-content-matt-brittin?CMP=share_btn_tw">Yvette Cooper</a> have requested), it has to define what is extremist.</p>
<p>The problem here is that Google’s approach to consulting with other institutions about how it operates <a href="http://deadline.com/2015/08/google-david-garrett-film-piracy-you-tube-edinburgh-1201509204/">is often spiky</a> – as their engagement with the ASA illustrates. If they only listen to what brands want, it leaves them very susceptible to the types of insidious biases that Brin and Page spoke about in their 1998 paper: see its definition of “<a href="https://support.google.com/youtube/answer/6162278?hl=en-GB">advertiser-friendly content</a>”. Might they start to direct viewers towards advertising-friendly content? They could, for example, push other content down their search results and recommended video lists. </p>
<p>So what does Google do next in the stand off with the big brands? Chief Business Officer <a href="https://www.bloomberg.com/news/articles/2017-04-03/google-updates-ads-polices-again-ramps-up-ai-to-curtail-youtube-crisis">Philip Schindler told Bloomberg</a> that the firm has improved its ability to flag offending videos and immediately disable ads. He also revealed that giant health products firm, Johnson & Johnson, has rolled back its boycott in some major markets. More broadly though, the solution is for Google to be far more transparent and engaged in the societies in which it operates. It needs to respond to brands and consumers but other social institutions as well. We all have to decide what is acceptable, not an algorithm, brands, politicians or individual YouTube viewers.</p><img src="https://counter.theconversation.com/content/75033/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Robert Cluley 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 search engine’s founders saw this one coming 20 years ago. So how should they react now?
Robert Cluley, Assistant Professor in Marketing, University of Nottingham
Licensed as Creative Commons – attribution, no derivatives.