AI
AI Takes a Seat at the Security Council as Altman and Amodei Brief the UN
On Wednesday, artificial intelligence took a seat at the table where the world discusses war and peace. Sam Altman of OpenAI, appearing in person, and Dario Amodei of Anthropic, joining remotely, briefed the 15 member UN Security Council on AI risks, alongside Hugging Face CEO Clément Delangue and researcher Yoshua Bengio, co chair of the UN's Independent International Scientific Panel on AI. France, holding the council's rotating presidency, convened the session, and France's Foreign Minister Barrot chaired it.
This was the council's first meeting focused specifically on safety risks from increasingly capable AI, including misalignment and a possible loss of human control. The council discussed AI in 2023 and again in 2024. Wednesday's briefing went further, placing AI under the council's peace and security agenda. That framing is the development that matters. Once an issue lands under the Security Council's maintenance of international peace and security agenda, the precedent exists for future action, including, eventually, binding resolutions.
The guest list crossed every fault line in the debate. Altman was expected to push international benchmarks for measuring AI capabilities and evaluating developer safeguards. Amodei has called for slowing AI development and continuing US chip export restrictions on China. China's DeepSeek and Moonshot were invited to make statements, with DeepSeek expected to participate, though founder Liang Wenfeng was expected to skip the in person seat. Secretary General António Guterres has said AI systems must remain under human direction, insight and control. Meanwhile President Trump told the General Assembly the same day that America would encourage AI, and keep restrictions off the table, placing the White House in direct tension with the governance requests coming from US labs themselves.
Industry arrived prepared. The Blueprint Alliance, bringing together Okta, CrowdStrike, AWS, Google Cloud and others, aims to create open standards and automated controls for managing autonomous AI agents. Gartner projects Fortune 500 companies could run more than 150,000 agents each by 2028, which explains why the standards conversation is happening now, before the agent population explodes.
The takeaway is diplomatic, and it favors the builders who cooperate early. When the Security Council starts taking briefings, the era of pure self governance is giving way to something shared, and the labs shaping the standards today will be the ones living comfortably under them tomorrow.
France is using its G20 presidency to push shared global principles on AI governance, alongside proposals for the UN to certify independent AI research institutes and launch a capacity building fund for developing countries. That capacity angle carries a libertarian warning worth hearing. Whoever sets the benchmarks gets to decide which AI projects are legitimate, and developing countries are being asked to trade sovereignty for access. The labs building open weights models have their own answer. Keep the frontier open, keep measurement honest, and let a thousand experiments run.
Quick answers
What is this story about?
On Wednesday, artificial intelligence took a seat at the table where the world discusses war and peace. Sam Altman of OpenAI, appearing in person, and Dario Amodei of Anthropic, joining remotely, briefed the 15 member UN Security Council on AI risks, alongside Hugging Face CEO Clément Delangue and researcher Yoshua Bengio, co chair of the UN's Independent International Scientific Panel on AI. France, holding the council's rotating presidency, convened the session, and France's Foreign Minister Barrot chaired it.
Why does this story matter?
France is using its G20 presidency to push shared global principles on AI governance, alongside proposals for the UN to certify independent AI research institutes and launch a capacity building fund for developing countries. That capacity angle carries a libertarian warning worth hearing. Whoever sets the benchmarks gets to decide which AI projects are legitimate, and developing countries are being asked to trade sovereignty for access. The labs building open weights models have their own answer. Keep the frontier open, keep measurement honest, and let a thousand experiments run.
Sources
New to crypto? Read the crypto glossary, browse frequent questions, read our story, or explore the story archive.