UN Security Council Hears Frontier CEOs Warn AI Could Escape Human Control
OpenAI and Anthropic leaders told the UN Security Council on 23 September 2026 that poorly managed frontier AI could threaten humanity, days after both firms shipped cheaper flagship models.
Classy aggregates publicly available material; we did not conduct a private interview.
Watch the public session
What changed
On 23 September 2026, Sam Altman of OpenAI and Dario Amodei of Anthropic addressed the United Nations Security Council in separate briefings during the 81st General Assembly session in New York. Yoshua Bengio, co chair of the UN International Independent Panel on AI, and Hugging Face chief executive Clément Delangue also spoke.
Altman spoke for roughly ten minutes. Associated Press reporting quotes him warning that humanity could lose control of the future to AI if governments fail to coordinate safeguards. Amodei told the council that if managed poorly, I even believe AI could be a risk to humanity as a whole, according to PBS NewsHour coverage.
The briefings landed one day after Anthropic released Claude Opus 5.5 and OpenAI released GPT 6 Sol and GPT 6 Luna on 22 September 2026.
Why it matters
Security council attention signals that AI risk is now framed as a geopolitical stability question. For enterprise buyers, public pleas for slowdown paired with same week model releases force a credibility test on vendor safety claims.
Altman urged the council to avoid doomerism, the Guardian reported, while still acknowledging existential concern threads. Bengio’s published remarks stressed an urgent mission for verifiable control of frontier systems.
Who is affected
National security staffs, UN member missions, enterprise AI governance committees, and insurers writing emerging technology policies inherit new public statements they can cite in briefings.
What to do next
Document which models your organization runs against the public statements each vendor made at the council. Require written answers on international testing participation before renewing enterprise agreements this quarter.
What to watch
Follow whether any Security Council member introduces a resolution with measurable compliance hooks, not only voluntary pledges.
Sources
- Primary. C SPAN, Artificial Intelligence CEOs Brief the U.N. Security Council on the Risks of AI (23 September 2026). Public video record of remarks.
- Secondary. PBS NewsHour, AI firm leaders tell UN Security Council that it could be a risk to all humanity (23 September 2026). Direct quotes from Altman and Amodei.
- Secondary. The Guardian, OpenAI’s Altman and Anthropic’s Amodei address UN security council (23 September 2026). Context on timing and governance language.