🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

The full transcript and video from our event with David Autor, Ioana Marinescu, Katya Klinova, & Anton Korinek on the MIT's Work for the Future Task Force results are now available. fhi.ox.ac.uk/david-autor/

Four new published papers from the Causal Incentives Working Group (a collaboration with Oxford, DeepMind and U of Toronto researchers), on incentive concepts, and how they can be applied to AI safety. In AAAI x2, AAMAS, and Synthese. fhi.ox.ac.uk/causal-incenti…

“Understanding the effectiveness of government interventions in Europe’s second wave of COVID-19” collaboration b/w FHI, Imperial College London, OATML_Oxford, University of Copenhagen, London School of Hygiene & Tropical Medicine, Cambridge University, The Crick Paper medrxiv.org/content/10.110… Summary of results x.com/sorenmind/stat…

Why did the US and USSR fail to agree on int'l control of nuclear weapons in 1944-1951? Waqar Zaidi and Allan Dafoe's recently published International Control of Powerful Technology: Lessons from the Baruch Plan for Nuclear Weapons. Paper here: ow.ly/9kag50EhU0l

What are the fundamental causal limits on how much of the universe we can observe or affect? Toby Ord recently published The Edges of Our Universe. Paper here: arxiv.org/abs/2104.01191

FHI won't be running summer fellowships this year, but our colleagues at the Stanford Existential Risks Initiative CISAC & on Bluesky @stanfordcisac.bsky.social have a similar programme, and some FHI researchers may supervise relevant projects. Applications due April 21st! tinyurl.com/seri21summer

How can we forecast and mitigate AI’s impacts on inequality? Join us on Monday for an in-depth conversation between Joseph E. Stiglitz and Anton Korinek about AI and inequality. Monday 3rd May 1700–1830 BST (0900–1030 PT, 1200–1330 EDT) Registration: crowdcast.io/e/Stiglitz

#FutureProofReport, a new report by Toby Ord and The Centre for Long-Term Resilience takes a look at how the UK Government can be better prepared for the next extreme risk event - whether a pandemic or otherwise. longtermresilience.org/futureproof

Join us on Monday for Katya Klinova and Stephanie Bell from Partnership on AI presenting on 'Redesigning AI for Shared Prosperity' Rob Seamans will join as discussant and Anton Korinek will moderate Mon 21st Jun 1700–1830 BST (0900–1030 PT, 1200–1330 EDT) Register: ow.ly/Bg8u50FcsU9

Join us in an hour for Katya Klinova and Stephanie Bell from Partnership on AI presenting on 'Redesigning AI for Shared Prosperity' Rob Seamans will join as discussant and Anton Korinek will moderate Register: ow.ly/Bg8u50FcsU9

What are the harms of epistemic fragmentation in online targeted advertising using ML? Silvia Milano recently coauthored 'Epistemic fragmentation poses a threat to the governance of online targeting' in Nature Machine Intelligence Read the full paper here: ow.ly/gQFX50FgUTj

By how much does community mask-wearing reduce COVID-19 transmission? A new paper coauthored by FHI researchers explores the relationship between mask-wearing and transmission Read the full paper here: ow.ly/fVBZ50FgTYU

Recordings and transcripts of our past two webinars are now available: ow.ly/qBQQ50FiAJT Redesigning AI for Shared Prosperity: An Agenda by Stephanie Bell and Katya Klinova and AI and Inequality: Joseph Stiglitz in discussion with Anton Korinek

Over the past years I have been deeply impressed by @DeepMind, both by its technical achievements and its leadership in the responsible development of AI. I am delighted to share that I have recently joined DeepMind, where I will continue my work on AI governance. 1/3

Today, we’re relaunching as a nonprofit. Our mission remains the same: We are building a global research community, dedicated to helping humanity transition to a world with advanced AI. But we’ll also be exploring new ways to expand our impact. governance.ai/post/the-centr…

New research from our AI safety team argues that human extinction is a likely outcome of running a sufficiently advanced artificial agent. x.com/Michael0515600…

The audiobook of The Precipice is on sale at a quarter the normal price for the next two weeks. audiobooks.co.uk/promotions/pro…

Here's a poster and video summary of recent work at AI Magazine! We found that under certain (very plausible) conditions, we can expect advanced artificial agents to intervene in the provision of rewards, with catastrophic consequences. 1/2