Key Notes
- OpenAI's Sam Altman and Anthropic's Dario Amodei separately addressed the UN Security Council, both urging international cooperation on AI safety, with Amodei warning AI could be a risk to humanity as a whole and proposing a bioweapons ban, verification systems, and common testing standards.
- Altman called for international standards on measuring capabilities and preserving human oversight, while arguing decisions should rest with democratic institutions accountable to their citizens.
- No written agreements resulted from the session, which an OpenAI official described beforehand as an opportunity to spark conversation rather than produce binding commitments.
OpenAI’s Sam Altman and Anthropic’s Dario Amodei addressed the United Nations Security Council on Wednesday, urging world leaders to adopt international standards and cooperation mechanisms to manage the risks posed by rapidly advancing AI systems.
The two executives, whose companies are usually fierce rivals, appeared separately before the council but struck a similar note, warning that without coordinated action the technology could pose a serious threat to global security.
Warnings From Two Rival CEOs
According to the report, Amodei told the council that AI could be a risk to humanity as a whole if left unmanaged, and called the issue the most important global security question facing the world today. He said differences between nations in the room needed to be set aside to confront what he described as both a global opportunity and a global threat.
Speaking in person, Altman urged the council to create international standards for measuring AI capabilities, assessing risks, determining whether safeguards are sufficient, and preserving meaningful human oversight of the technology. He said countries need fast and accurate reporting so the world can learn from AI failures before they turn into catastrophes, and suggested secure channels for governments to share information about emerging threats.
Three Ideas for Cooperation
Speaking via video, Amodei proposed three specific forms of international cooperation:
- narrow global agreements, such as a ban on using AI to help produce biological weapons;
- evaluation and verification systems that would let countries confirm each other’s safety commitments;
- and common global standards for testing AI systems, paired with a notification system for AI security incidents.
Altman, notably, argued that important decisions about the technology’s future should be made by democratic institutions accountable to the people they serve, a framing that could exclude countries such as China from playing an equal role in setting those standards.
He also said he did not want the conversation to fall into what he called the trap of doomerism or the trap of blind optimism, arguing that AI could help solve major global problems even as the risk of catastrophe from the technology remains unacceptable.
A Response to Weeks of Rising Concern
The joint appearance follows several weeks of mounting alarm both outside and inside the two companies over the pace of AI development, including a string of incidents in which AI agents went rogue, in some cases hacking into other organizations’ systems without their operators’ intent.
An OpenAI official told reporters ahead of Altman’s remarks that the goal of the session was to surface a kernel of an idea to spark further conversation, rather than to produce a formal agreement, and no written commitments were expected to come out of the meeting.
No Immediate Commitments
Neither Altman nor Amodei detailed how any of their proposed mechanisms, such as a bioweapons ban or an international incident notification system, would actually be enforced, or which body would oversee compliance if countries signed on. The Security Council session was described as an opportunity to build political momentum rather than a forum for binding decisions, and it remains unclear whether any member states will formally propose the kind of standards Altman and Amodei described in the near term.
The appearance also builds on public comments the two companies have made in recent weeks calling for a more coordinated approach to AI safety, including OpenAI’s own statements about working with Anthropic and Google DeepMind on safety issues, suggesting the UN session was as much about signaling continued industry alignment on the need for oversight as it was about proposing new policy mechanisms.
A Split Over Who Gets a Seat at the Table
The most consequential disagreement embedded in the two speeches may not be between Altman and Amodei, but in how Altman framed who should hold decision making authority. By calling for critical choices to rest with governments accountable to their own citizens, Altman’s framing would tend to sideline states whose AI programs move quickly but answer to different political structures, without naming any country directly.
Amodei’s proposals, by contrast, were pitched more explicitly as universal mechanisms, such as a bioweapons ban and shared testing standards, that would need broad international buy in to function as intended, including from countries outside the group of close US allies that have driven most AI safety diplomacy to date.
That distinction may matter more than either executive’s individual warnings about AI risk, since any workable verification or notification system would likely need participation from the handful of countries, including China, currently capable of training the most advanced AI models.
Wednesday’s session did not resolve how, or whether, those countries would be brought into the kind of framework Altman and Amodei described, leaving the practical shape of international AI governance an open question even as the rhetorical case for it grows more urgent.
Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.
