<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:media="http://search.yahoo.com/mrss/">
<channel>
<title>deepmind — GZERO Media</title>
<link>https://www.gzeromedia.com/tag/deepmind</link>
<description>The latest news and analysis from GZERO Media.</description>
<language>en</language>
<copyright>© GZERO Media</copyright>
<atom:link href="https://www.gzeromedia.com/feeds/tag/deepmind.xml" rel="self" type="application/rss+xml" />
<lastBuildDate>Tue, 20 May 2025 11:30:02 +0000</lastBuildDate>
<item>
<title>An OpenAI insider warns of the reckless race to AI dominance</title>
<link>https://www.gzeromedia.com/video/gzero-world-clips/an-openai-insider-warns-of-the-reckless-race-to-ai-dominance</link>
<guid isPermaLink="false">cmvrg6jhmly6g07lqhqgwvdb0</guid>
<pubDate>Tue, 20 May 2025 11:30:02 +0000</pubDate>
<dc:creator>GZERO Staff</dc:creator>
<category>deepmind</category>
<category>ai</category>
<category>ian bremmer</category>
<category>artificial general intelligence</category>
<category>daniel kokotajlo</category>
<category>gzero world</category>
<category>openai</category>
<category>GZERO World Highlights</category>
<description>Former OpenAI researcher Daniel Kokotajlo warns that tech companies are racing toward superintelligent AI while ignoring safety issues or democratic oversight — putting humanity at serious risk.</description>
<media:content url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrddbgxe60h07lqyaiou5uz" type="image/jpeg" medium="image" width="1280" height="720">
<media:credit>youtu.be</media:credit>
<media:text>- YouTube</media:text>
</media:content>
<media:thumbnail url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrddbgxe60h07lqyaiou5uz" />
<content:encoded><![CDATA[<p>Are AI companies being reckless and ignoring safety concerns in the race to develop superintelligence? On GZERO World, <strong>Ian Bremmer</strong> is joined by former OpenAI whistleblower and executive director of the AI Futures Project, <strong>Daniel Kokotajl</strong><strong>o</strong>, to discuss new developments in artificial intelligence and his concerns that big tech companies like OpenAI and DeepMind are too focused on beating each other to create new, powerful AI systems and not focused enough on safety guardrails, oversight, and existential risk. Kokotajlo left OpenAI last year over deep concerns about the direction of its AI development and argues tech companies are dangerously unprepared for the arrival of superintelligent AI. If he’s right, humanity is barreling toward an era of unprecedented power without a safety net, one where the future of AI is decided not by careful planning, but by who gets there first.</p><p>“OpenAI and other companies are just not giving these issues the investment they need,” Kokotajlo warns, “We need to make sure that the control over the army of superintelligences is not something one man or one tiny group of people gets to have.”</p><h5><a href="https://www.gzeromedia.com/video/gzero-world-with-ian-bremmer/">GZERO World with Ian Bremmer,</a> the award-winning weekly global affairs series, airs nationwide on US public television stations (check local listings).</h5><p><strong>New digital episodes of GZERO World are released every Monday on YouTube. Don't miss an episode: </strong><a rel="noopener noreferrer" target="_blank" href="http://www.youtube.com/@GZEROMedia"><strong>subscribe to GZERO's YouTube channel&nbsp;</strong></a><strong>and turn on notifications (🔔).GZERO World with Ian Bremmer airs on US public television weekly - check local listings.</strong></p>]]></content:encoded>
</item>
<item>
<title>Google’s own mathlete</title>
<link>https://www.gzeromedia.com/gzero-ai/googles-own-mathlete</link>
<guid isPermaLink="false">cmvrft002kq4t07lqmrikto7c</guid>
<pubDate>Tue, 30 Jul 2024 16:11:36 +0000</pubDate>
<dc:creator>Scott Nover</dc:creator>
<category>math</category>
<category>deepmind</category>
<category>mathematical olympiad</category>
<category>ai</category>
<category>google</category>
<category>science-tech</category>
<category>GZERO AI</category>
<description>Google’s best artificial intelligence system is now doing math at a high school level. Okay, well, a really, really, really good high school level.</description>
<media:content url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrcqu91cab707myvsih80cu" type="image/jpeg" medium="image" width="1190" height="676">
<media:credit>REUTERS/Gonzalo Fuentes/File Photo</media:credit>
<media:text>People silhouetted in front of a Google logo during the inauguration of a new hub in France dedicated to the artificial intelligence sector at the Google France headquarters in Paris, France.</media:text>
</media:content>
<media:thumbnail url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrcqu91cab707myvsih80cu" />
<content:encoded><![CDATA[<p>Google’s best artificial intelligence system is now doing math at a high school level. Okay, well, a <em>really, really, really</em> good high school level.</p><p>A pair of Google DeepMind models tried to solve the <a target="_blank" href="https://deepmind.google/discover/blog/ai-solves-imo-problems-at-silver-medal-level/"><u>six problems</u></a> posed to teenage math whizzes at the International Mathematical Olympiad and came away with a silver medal performance, the first time an AI model has been medal-worthy.<br></p><p>The models, called AlphaProof and AlphaGeometry 2, were given unlimited time while students only got 4.5 hours per test, but the fact that it could reason through these complicated advanced algebra, geometry, and number theory problems at all was an achievement. It did flunk two combinatorics problems, which have to do with counting and arranging large quantities.</p><p><strong>David Silver</strong>, Google DeepMind’s vice president of reinforcement learning, said this <a rel="noopener noreferrer" target="_blank" href="https://www.nytimes.com/2024/07/25/science/ai-math-alphaproof-deepmind.html"><u>represents</u></a> a “step-change in the history of mathematics.” Hopefully, he said, it’s not just a big moment for the Olympiad, but also “represents the point at which we went from computers only being able to prove very, very simple things toward computers being able to prove things that humans can’t. In the future, Google hopes that its models can work hand-in-hand with human mathematicians to solve the hardest problems in the field, though first, it seems, they’ll have to conquer the Olympiad.</p>]]></content:encoded>
</item>
<item>
<title>Hard Numbers: Stolen art, mathletes, DeepMind defection, Antitrust tussle, Labor shortages</title>
<link>https://www.gzeromedia.com/gzero-ai/hard-numbers-stolen-art-mathletes-deepmind-defection-antitrust-tussle-labor-shortages</link>
<guid isPermaLink="false">cmvrfjhu5judo07lq7k9k4l0a</guid>
<pubDate>Tue, 23 Jan 2024 18:04:59 +0000</pubDate>
<dc:creator>Scott Nover</dc:creator>
<category>science-tech</category>
<category>deepmind</category>
<category>midjourney</category>
<category>ai</category>
<category>google</category>
<category>japan</category>
<category>jobs</category>
<category>GZERO AI</category>
<description>16,000: A list of 16,000 names of artists that Midjourney allegedly trained its image generation model on has spread on the internet after leaking from an online court filing.</description>
<media:content url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrcbbgtc5oj07lqm22aow74" type="image/jpeg" medium="image" width="1200" height="900">
<media:credit>IMAGO/Action Pictures via Reuters Connect</media:credit>
<media:text>Midjourney artificial intelligence on a computer screen.</media:text>
</media:content>
<media:thumbnail url="https://us-west-2.graphassets.com/cmrcgvnh6064507n8hhitfmjj/cmvrcbbgtc5oj07lqm22aow74" />
<content:encoded><![CDATA[<p><strong>16,000:&nbsp;</strong>A list of <a target="_blank" href="https://www.theguardian.com/technology/2024/jan/21/we-need-to-come-together-british-artists-team-up-to-fight-ai-image-generating-software">16,000</a> names of artists that Midjourney allegedly trained its image generation model on has spread on the internet after leaking from an online court filing. The list purports to show that Midjourney trained its model on the works of artists like Frida Kahlo and Damien Hirst. The list is an exhibit from an ongoing class-action lawsuit against the AI company and was accidentally posted online before being removed.</p><p><strong>25:&nbsp;</strong>A math-focused AI system is almost at the level of our species’ most advanced teenage mathletes. The model called AlphaGeometry reportedly scored a <a target="_blank" href="https://www.nytimes.com/2024/01/17/science/ai-computers-mathematics-olympiad.html">25</a> out of 30 on the International Mathematical Olympiad, the premier high school math competition. The average gold medalist in the competition has scored a 25.9.<br></p><p><strong>220 million:&nbsp;</strong>A pair of Google DeepMind engineers are reportedly ready to leave the company, in talks to raise <a rel="noopener noreferrer" target="_blank" href="https://www.bloomberg.com/news/articles/2024-01-19/google-deepmind-ai-scientists-in-talks-to-leave-for-french-stealth-startup?embedded-checkout=true">$220 million</a> to build a new large language model. The company, which would be based in France, is tentatively called Holistic.</p><p><strong>2:&nbsp;</strong>The US government's effort to investigate Microsoft’s $13 billion investment in OpenAI is on hold. The problem: <a rel="noopener noreferrer" target="_blank" href="https://www.politico.com/news/2024/01/19/doj-ftc-microsoft-openai-antitrust-00136624">Two different</a> agencies can’t decide on who should take the lead. This is the problem that’s arisen between the country’s two antitrust authorities, the Department of Justice and the Federal Trade Commission. Neither party seems ready to relinquish the job to the other.</p><p><strong>11 million:</strong> Japan is estimated to face a labor shortage of <a rel="noopener noreferrer" target="_blank" href="https://www.ft.com/content/ad850ad2-6752-4ca7-99f6-4b947d0b741e">11 million</a> people by 2040 as its aging population peaks and exits the workforce. While AI and robots may displace human labor in some economies, they might also help others — such as Japan — avert crises in industries such as construction and retail.</p>]]></content:encoded>
</item>
</channel>
</rss>