- Published on
UN Security Council Will Get Advice on AI Risks From Tech Giants Building It
- Authors
- Name
- DeepHorus
The UN Security Council will hear directly from the technology leaders building artificial intelligence on Wednesday, as it assesses the risks posed by advanced AI systems. OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, AI safety researcher Yoshua Bengio and Hugging Face CEO Clément Delangue are scheduled to brief the 15-member body. Chinese firms DeepSeek and Moonshot have also been invited to make statements. The session falls two days before President Donald Trump meets Chinese President Xi Jinping in Washington on September 24.
France organized the briefing and holds the council's rotating presidency this month. Foreign Minister Jean-Noël Barrot will chair the meeting. Its concept note reportedly centers on AI systems slipping beyond human control. The council, which can authorize sanctions and military action, is hearing from tech executives instead of generals.
Discussion is expected to cover AI that could eventually run cyberattacks, sway elections or design weapons without a person in the loop. DeepSeek drew attention in January 2025 after its cut-price model wiped roughly $600 billion from Nvidia's market cap. Moonshot is the developer of the Kimi models. Amodei will participate remotely.
Amodei published an essay, "We Must Pace the Frontier," on September 12, arguing the industry should deliberately slow development of the most advanced frontier models. He said Anthropic would grant outside evaluators permanent, staff-level access to verify its safety promises. Sam Altman and Elon Musk both said they agreed within hours.
Amodei called China "the autocratic country with by far the most advanced AI capabilities" and urged Washington to keep restricting advanced chip exports to Beijing. China's Foreign Ministry called the essay counterproductive, and state media branded it a "Cold War playbook."
President Trump separately dismissed AI-takeover warnings as a "hoax" tied to the "Radical Left" and accused Amodei of wrongdoing without offering evidence. King Charles later hosted executives from Anthropic, OpenAI, Nvidia and Google DeepMind at a Scottish estate summit that produced no publicly known agreements.
Wednesday's session is not expected to yield a binding deal. AI safety is reportedly on the agenda for the September 24 Trump-Xi summit, alongside trade.
