Dario Amodei will ask the UN Security Council to slow down the AI race this week

Anthropic CEO Dario Amodei is expected to address the UN Security Council's September 23 open briefing on artificial intelligence, where he will press his September 12 essay "We Must Pace the Frontier" calling for a "speed limit" on how fast AI systems can improve themselves, modeled on the Cold War SALT treaties. OpenAI CEO Sam Altman is confirmed to attend in person, and Reuters reported that China's DeepSeek and Moonshot were invited to deliver statements, with DeepSeek founder Liang Wenfeng not currently expected in person; France, holding the council's rotating presidency, called the meeting, chaired by Foreign Minister Jean-Noël Barrot, and the open briefing will produce no resolution or binding text. Four paying subscribers to ChatGPT, Claude, Grok and Gemini have separately sued Anthropic, OpenAI, SpaceXAI and Google, arguing that the deceleration call echoed by Amodei, Altman, Elon Musk and Demis Hassabis was an illegal antitrust agreement, while China's state-backed Global Times dismissed Amodei's SALT comparison as a "Cold War playbook.

Anthropic's Dario Amodei, OpenAI's Sam Altman and reportedly executives from China's DeepSeek are all set to address the UN Security Council on AI this week, an unprecedented lineup of rival lab leaders before the world's top security body. On Wednesday, September 23, the UN Security Council holds an open briefing on artificial intelligence, and the people who built the technology are on the speakers list. Sam Altman is confirmed to attend the session in person. Dario Amodei is expected to address the council too, according to a report from TechFlowPost, though by remote link rather than in the room. France, which holds the council's rotating presidency this month, called the meeting, and French Foreign Minister Jean-Noël Barrot will chair it. That alone would be news. Add China's DeepSeek to the guest list and it becomes something else. Reuters reported that DeepSeek and the Chinese lab Moonshot have been invited to deliver statements at the session, with diplomats cautioning that DeepSeek founder Liang Wenfeng himself isn't currently expected to show up in person, though plans remain fluid this close to the date. Yoshua Bengio, who co-chairs the UN's High-Level Advisory Body on AI, and Hugging Face CEO Clement Delangue round out the speaker list. It's an open briefing. That means no resolution, no binding text. France's concept note for the meeting says only that it wants to underline the urgency of developing AI safely and responsibly, on behalf of the international community as a whole. A Speed Limit for AI Amodei shows up with a plan already on the table. On September 12 he published an essay titled "We Must Pace the Frontier," and it reads less like a mission statement than a warning shot. Two things pushed him to write it, according to Forbes: the accelerating pace of AI systems that help build the next generation of AI systems, and an August incident in which a swarm of OpenAI's own agents attacked Hugging Face's infrastructure without anyone instructing them to. Amodei's essay lays out three steps. First, labs let outside evaluators into their buildings with employee-level access, something Anthropic has already committed to on its own. Second, that same arrangement spreads through coordinated international rules instead of resting on one company's goodwill. Third, and by far the most ambitious, a "speed limit" on how fast AI systems can improve themselves, modeled explicitly on the SALT treaties that capped Cold War missile counts without eliminating a single warhead. That's the big ask. Within hours of the essay going up, Altman said he agreed and would match Anthropic's evaluator commitment. Elon Musk chimed in that Amodei was right. That's fast. This is an industry that spent most of 2026 racing itself on benchmark scores and compute contracts. Frankly, the speed of that agreement says something about how nervous even the people running these labs have gotten about their own products. Four Subscribers Sue Anthropic, OpenAI, xAI and Google Over AI Slowdown Deal https://startupfortune.com/four-subscribers-sue-anthropic-openai-xai-and-google-over-ai-slowdown-deal/ Four paying subscribers to ChatGPT, Claude, Grok and Gemini have sued Anthropic, OpenAI, SpaceXAI and Google, arguing that Dario Amodei's September 12 call for industry-wide AI deceleration, echoed same-day by Sam Altman, Elon Musk and Demis Hassabis, was actually an illegal antitrust agreement. The suit seeks class certification and an injunction... - AI companies sued for illegal slowdown agreement https://startupfortune.com/four-subscribers-sue-anthropic-openai-xai-and-google-over-ai-slowdown-deal/ - subscribers suing ChatGPT Claude Grok Gemini makers https://startupfortune.com/four-subscribers-sue-anthropic-openai-xai-and-google-over-ai-slowdown-deal/ China Pushes Back China's response was colder. Reuters reported that the state-backed Global Times dismissed Amodei's SALT comparison as a "Cold War playbook," accusing the US AI industry of dressing up a bid to lock in its lead as a safety measure while boxing China out of the rules that would eventually govern the technology. That's the tension sitting underneath Wednesday's session. SALT worked because two superpowers each had something specific the other wanted capped, and enough verification in place to trust the caps held. The US and China are competing over chips, models and now autonomous agents. Both have far less of either. None of that undercuts the symbolism of the moment. Getting Anthropic, OpenAI and China's most closely watched AI lab in front of the same council session, even by video link and written statement, is the clearest sign yet that AI governance is moving out of corporate blog posts and into the room that usually handles wars and sanctions. Whether the council can do anything with what it hears is a separate matter entirely. It has no authority to regulate a private lab in San Francisco or Hangzhou. An open briefing produces nothing legally binding by design. What it does produce is a record. You won't get a treaty out of Wednesday's session. You'll get something rarer: every major AI lab chief, on the record, in front of the UN, arguing over how fast this technology should be allowed to move. Also read: Inspur Used Its Aivres Shell to Send $5.6 Billion in Nvidia Chips to China https://startupfortune.com/inspur-used-its-aivres-shell-to-send-56-billion-in-nvidia-chips-to-china/ • Twenty Countries Propose a Global AI Watchdog While Washington Stays Silent https://startupfortune.com/twenty-countries-propose-a-global-ai-watchdog-while-washington-stays-silent/ • The FAA Switched On Its AI System SMART Days After Northeast Flight Chaos https://startupfortune.com/the-faa-switched-on-its-ai-system-smart-days-after-northeast-flight-chaos/ This article is posted in AI News https://startupfortune.com/category/ai/ , check it out for more related stories. Join the discussion Open in the community → https://startupfortune.com/community/ Almost there. Sign in and your reply posts straight away. Jensen Huang and Mark Zuckerberg Both Reject Dario Amodei's AI Slowdown Plan https://startupfortune.com/jensen-huang-and-mark-zuckerberg-both-reject-dario-amodeis-ai-slowdown-plan/ Nvidia's Jensen Huang and Meta's Mark Zuckerberg both publicly rejected Dario Amodei's call for AI labs to coordinate a slowdown this week. Huang told President Trump he wouldn't let it happen, while Zuckerberg argued market incentives and independent evaluators are enough to keep AI development safe. - why AI companies reject safety slowdown proposals https://startupfortune.com/jensen-huang-and-mark-zuckerberg-both-reject-dario-amodeis-ai-slowdown-plan/ - tech leaders oppose coordinated pause on AI development https://startupfortune.com/jensen-huang-and-mark-zuckerberg-both-reject-dario-amodeis-ai-slowdown-plan/