Welcome to the Future of Life Institute newsletter. Every month, we bring 43,000+ subscribers the latest news on how emerging technologies are transforming our world.

If you've found this newsletter helpful, why not tell your friends, family, and colleagues to subscribe?

Today's newsletter is a 10-minute read. Some of what we cover this month:

• 🇰🇷 Reflections on the AI Seoul Summit

• 📰 OpenAI news

• ✅ The EU AI Act is (officially, finally) final!

• 🗺️ Results from the Foresight Institute’s worldbuilding program

• ⚖️ Governance updates from around the world

This month, the governments of the Republic of Korea and the UK cohosted the AI Seoul Summit, a follow-up to the UK AI Safety Summit held in November. The Summit brought together states, civil society, and industry to expand on discussions from the UK Summit, focused primarily on establishing international multi-stakeholder cooperation on safe AI innovation.

In advance of the Summit, 16 tech companies agreed on a voluntary set of AI safety commitments, but as FLI’s Director of Policy Mark Brakel told Euronews, voluntary commitments don’t go far enough - “goodwill alone is not sufficient”. We published our own recommendations to seize the unique opportunity for progress presented by the Summit, which you can read here.

The Summit concluded with a declaration signed by 27 states and the European Union, reaffirming their commitment to “collaborative international approaches” that harness AI’s benefits while mitigating its risks.

FLI President Max Tegmark was honoured to attend the Summit, with his comments on the Summit and his hopes for it featured in Reuters, TechTimes, and Fox.

Max spoke to The Guardian as well from the Summit, calling out Big Tech's work to shift attention away from existential risk presented by AI, in a distract-and-delay tactic intended to stave off meaningful regulation. Likening these efforts and industry lobbying to those employed for decades by the tobacco industry, Max again called for policymakers to take AI’s catastrophic risks seriously, and understand that we need to - and can - address both ongoing harms and escalating risks from AI.

FLI AI Safety Summit Lead Imane Bello also attended, and summarized her reflections on the Summit on X:

OpenAI’s safety team has been generating a lot of attention recently, following several notable departures over the past few months.

To recap: researchers Ilya Sutskever and Jan Leike, both of whom led OpenAI’s former Superalignment team - which focused on existential risks presented by AI - resigned within hours of each other earlier this month. At least six other departures have made news since the start of the year, with four of those employees having worked on the Superalignment team or AI governance in general.

Leike posted a thread on X detailing his reasons for leaving the company:

Earlier this week, OpenAI announced that they’re training a new “frontier model”, anticipating “the resulting systems to bring us to the next level of capabilities on our path to AGI”. They also announced the launch of a new Safety and Security Committee to assess and further develop OpenAI’s safety processes.

Also this week, former OpenAI board members Helen Toner and Tasha McCauley published an article in The Economist, arguing that the public sector and “prudent regulation” must be part of AI development.

• The EU AI Act is (finally) final! On 21 May, the Act - the world’s first comprehensive AI law - received approval from the European Council, marking the last hurdle it had to clear before becoming law. Implementation will begin soon, with it mostly coming into place over the next two years.

  • The EU AI Office has also just announced its structure, with current European Commission Director for AI and Digital Industry Lucilla Sioli formally leading it.

  • Be sure to check out FLI EU Research Lead Risto Uuk’s dedicated EU AI Act Newsletter for more details and analyses.

• The UK AI Safety Institute has announced a new £8.5 million grants programme to support research furthering systemic AI safety. Learn more here; applications to open shortly.

• The U.S. Bipartisan Senate AI Working Group unveiled its long-awaited Senate AI Roadmap, outlining the current status of policy proposals intended to support safe and responsible AI innovation, and potential next steps.

  • FLI’s Max Tegmark responded to the Roadmap, praising “this important step forward” while urging for "more action as soon as possible”.

• As the AI Seoul Summit kicked off, the U.S. Department of Commerce released their strategic vision on AI safety. They also announced the formation of a new international AI safety alliance.

  • The alliance will “strengthen and expand on” U.S. AI safety collaborations with the UK, Japan, Canada, Singapore, and the EU as they work to “promote safe, secure, and trustworthy artificial intelligence systems for people around the world”.