Altman Urges UN Security Council to Set Global AI Standards

OpenAI CEO Sam Altman took the case for coordinated global oversight of artificial intelligence to the United Nations Security Council on Wednesday, warning that increasingly capable and autonomous systems could outpace the institutions meant to govern them.

Addressing the council in New York, Altman said the pace of progress in recent weeks and months had made the timeline feel compressed, sharpening both the promise and the risks of the technology. He framed the moment as a fork in the road: AI could usher in what he described as a new renaissance of creativity and discovery, or resemble an industrial revolution marked by upheaval and disarray.

He pointed to the danger that AI systems, as they grow more capable, could move faster than institutions, concentrate power in too few hands, or make decisions that people no longer understand or control. That concern, he said, becomes especially significant as the field approaches systems capable of recursive self-improvement, where building AI becomes more automated and progress could accelerate rapidly. Such a prospect, he said, calls for extreme care.

Altman argued that the most consequential decisions about AI cannot rest with laboratories in San Francisco alone and must instead be shaped through democratic processes and by governments accountable to their citizens. At the international level, he said, cooperation is essential, recalling periods in history when competing countries set aside differences to confront a powerful new technology.

He called for a mechanism for complementary national and international frontier AI standards covering how capabilities are measured, how risks are assessed, whether safeguards are adequate, and how meaningful human oversight is preserved as systems become more autonomous. Common standards, he said, would let countries compare evidence, verify compliance and share a common language about what is happening.

Altman also pressed for accurate and speedy incident reporting, with classification and reporting protocols, so that the world can learn from failures before they become catastrophes, and for secure channels linking governments, critical infrastructure operators and technical experts to share emerging vulnerabilities and threats.

He cautioned that such standards should not entrench incumbents or favour one business model over another, and should accommodate open and closed model developers, new entrants and established labs, as well as national authority and international cooperation. Each government, he said, should decide how to incorporate the standards into its own legal system, but agreement on what constitutes good evidence, good safeguards and good oversight would benefit all.

Acknowledging that OpenAI does not have all the answers, Altman said the company is open to better ideas and ready to work with governments. He described the choice ahead as between a world where power is concentrated and decisions are opaque, and one where AI works for people because democratic institutions helped determine how it develops.

The session brought together heads of major artificial intelligence companies and Security Council members to consider whether the growing reach of AI should be controlled, and if so, by whom.