OpenAI wants Washington, not a patchwork of competing national regimes, to set the terms for how the world governs frontier AI.
OpenAI called on the United States on Monday to lead an international effort to set technical standards for frontier AI. The request, made in a company blog post, specifically named recursive self-improvement, AI systems that can enhance their own capabilities with limited human input, as a priority area for oversight.
The company wants a formal mechanism linking national and international standards, plus shared measurement tools and incident reporting protocols, so governments can act on AI risks collectively rather than in isolation.
The post was published as world leaders gathered in New York for the UN General Assembly, days after several AI industry figures, including OpenAI chief executive Sam Altman, called for a coordinated slowdown in AI development over fears systems could soon improve beyond human control.
"Leading now will determine whether the United States shapes the global AI framework," OpenAI wrote in the post.
OpenAI Defines "Pacing"
The blog post also picked up the industry term "pacing," the idea of deliberately managing how fast AI capabilities advance.
OpenAI defined pacing as making sure alignment research and safety deployment keep up with capability gains, rather than hitting any fixed development speed. Alignment refers to keeping AI systems' behaviour consistent with human values as they grow more capable.
Altman to Brief the Security Council
Altman is due to brief the UN Security Council's 15 members on Wednesday on AI and international security. France, which holds the Council presidency this month, convened the meeting; French Minister for Europe and Foreign Affairs Jean-Noël Barrot will chair it.
An OpenAI spokesperson said Altman's remarks would focus on international coordination and shared safety standards, alongside the steps the company is taking to ensure AI benefits people globally.
The Council first took up AI risk in 2023, when China warned against the technology becoming a "runaway horse" and the US cautioned against its use for censorship or repression. UN Secretary-General António Guterres has since called for coordinated global action, arguing that national measures alone won't be enough.
Anthropic chief executive Dario Amodei and Google DeepMind's Demis Hassabis have separately raised similar concerns: Amodei has argued the industry needs to slow how quickly it improves model capabilities, while Hassabis has called for a US-led global regulatory body.
Washington Isn't Convinced
Not everyone in Trump's administration shares that urgency. The White House has rejected calls to slow AI development, wary that a pause would hand China room to close the gap in a technology it treats as central to national and economic security.
Trump said earlier this month that the US already has sufficient guardrails in place, and argued that doubt cast on AI development would benefit China. Trump is scheduled to meet Chinese President Xi Jinping in Washington this week, with AI expected to come up in the talks.
Comments ( 0 )