OpenAI, Anthropic and Hugging Face CEOs brief UN Security Council on AI

By Splitfeed EditorialLast updated Sep 26, 2026, 21:10 UTC

On September 23, 2026, the UN Security Council convened its 10228th meeting focused on artificial intelligence and international security. Sam Altman of OpenAI and Dario Amodei of Anthropic addressed the council alongside representatives from Hugging Face and DeepSeek, making it a rare moment where leading AI developers from both the United States and China appeared together before the body. Amodei pledged during his address to slow down AI development, a commitment that stands in some tension with Anthropic's reported sevenfold revenue growth this year. Both Altman and Amodei told the council that no single nation or company should control AI. France formally flagged AI as a threat to children and global security ahead of the session, the UK Foreign Secretary addressed the council, and the United States delivered a formal intervention.

The meeting carries weight beyond its symbolic optics. The Security Council is not a regulatory body and has no direct mechanism to impose rules on private AI developers or member states, so the gathering raises genuine questions about what existing international institutions can actually do with the technology. Amodei's pledge to slow development is notable because Anthropic has simultaneously been scaling aggressively, which puts the sincerity and enforceability of that commitment in dispute. The inclusion of DeepSeek, a Chinese firm, alongside American companies also signals that at least some effort is being made to frame AI governance as a multilateral concern rather than a purely Western one, though how durable that framing proves remains an open question.

None of the three camps have published direct reactions to this specific meeting yet. The Pro-AI camp would typically argue that industry leaders engaging with international institutions demonstrates responsible self-governance and that voluntary commitments from developers are a more agile response than slow-moving treaty processes. The Anti-AI camp would typically contend that letting the same companies driving rapid AI deployment set the terms of any governance conversation is a conflict of interest, and that a body without enforcement power produces little more than public relations cover. The Middle Ground camp would typically welcome the multilateral format while warning that any meaningful framework needs binding mechanisms and participation from states beyond those already represented in the room.

The most concrete thing to watch is whether the session produces any formal Security Council resolution, statement or working group mandate, and whether Amodei's pledge to slow development is accompanied by any measurable benchmark or third-party verification process. A follow-up session or a report from the Secretary-General's office on AI governance could clarify how seriously member states intend to pursue the issue through the UN system.