AI’s Builders Tell the UN: Regulate Us

OpenAI CEO Sam Altman has taken an unusual message to the United Nations: governments, not technology companies, must ultimately set the rules governing increasingly powerful artificial intelligence.

Altman addressed the 15-member UN Security Council on Wednesday, September 23, as world leaders grapple with how to manage technology advancing faster than many of the institutions expected to regulate it.

His appearance placed one of the world's most prominent AI executives before the UN body charged with maintaining international peace and security — not simply to promote artificial intelligence, but to warn of what could happen if its development escapes effective human oversight.

Altman presented AI as carrying enormous potential alongside potentially severe risks. He said increasingly capable systems could accelerate scientific discovery, improve healthcare and education, strengthen economies and give individuals powerful new tools.

But he also identified two dangers that he argued must be avoided: humans losing control of increasingly autonomous AI systems, and control of powerful AI becoming concentrated among a handful of companies, individuals or countries.

Altman said competitive pressure between technology companies should not justify taking excessive risks and indicated that OpenAI has previously slowed development when necessary and would do so again.

He argued that developers should not train increasingly powerful models unless they can make a strong case that the systems will remain under human control.

But one of the most significant parts of his address concerned who should ultimately govern the technology.

Altman argued that private AI laboratories should not substitute themselves for democratic institutions. Governments accountable to their populations, he said, must help determine the rules under which increasingly powerful systems operate.

At the international level, he called for cooperation between countries on common standards for advanced AI.

Those standards would cover how AI capabilities and risks are measured, whether safeguards are adequate, how meaningful human oversight is preserved and how countries verify compliance.

Altman also advocated rapid reporting of AI-related incidents and secure channels through which governments, critical infrastructure operators and technical experts could share information about emerging vulnerabilities and threats.

The intervention is notable because OpenAI is itself one of the companies driving the rapid development of the technology for which its chief executive is seeking stronger governance.

Altman was not alone.