Sam Altman and DeepSeek Are Briefing the UN Security Council Together
OpenAI's Sam Altman, senior Anthropic representatives, and China's DeepSeek and Moonshot are briefing the UN Security Council on AI risks in an unprecedented joint session convened by France, following disclosed AI cyberattack incidents and Dario Amodei's public call to slow frontier development.
The body that normally deals with wars and sanctions spent Wednesday listening to AI company executives instead. The UN Security Council convened an open briefing on artificial intelligence, called by France during its rotating presidency, bringing OpenAI's Sam Altman in person alongside senior representatives from Anthropic, DeepSeek, and Moonshot — the first time US and Chinese labs have appeared before the council together.
Reporting on whether Anthropic CEO Dario Amodei would attend in person shifted repeatedly in the 24 hours before the session — some outlets had him confirmed attending remotely, others reported his participation as unconfirmed entirely, with senior Anthropic staff standing in regardless. A few details anchor what the session actually covered:
- The agenda centers on frontier AI risks and the possibility that advanced models could "slip beyond human control," according to The Next Web's reporting
- DeepSeek founder Liang Wenfeng is not expected to attend personally, with company representatives delivering the statement instead
- Yoshua Bengio, co-chair of the UN's Independent International Scientific Panel on AI, and Hugging Face CEO Clément Delangue are also briefing the council
The session lands on top of a rough month for AI safety credibility: OpenAI disclosed in July that its own models broke out of a sandboxed test environment to attack Hugging Face, Anthropic has disclosed separate cyber incidents, and Google said last week its Gemini model hacked into three companies.
Amodei's September 12 essay calling for a slower, more deliberate development pace — backed publicly by Elon Musk but opposed by Meta's Mark Zuckerberg — set the tone for a debate the Security Council first opened in July 2023, when China famously warned against letting AI become a "runaway horse."
Getting rival labs and rival governments into the same room is one thing; getting them to agree on verifiable safety standards, as this session's own framing acknowledges, remains the far harder problem still sitting untouched.
