White House Turns Down OpenAI and Anthropic Push for Global AI Safety Standards

The United States has rejected appeals from OpenAI and Anthropic to throw its weight behind global standards for artificial intelligence, according to reporting from the BBC. The decision leaves the two leading American labs without the federal backing they had sought as they push for shared international rules on how the most capable AI systems are built and released. The rejection lands at a moment when governments in Europe and Asia are moving ahead with their own AI statutes, and when US lawmakers have yet to pass comprehensive federal legislation of their own.

OpenAI and Anthropic had argued that a single set of cross-border expectations would reduce the patchwork of rules the companies face and give them clearer ground to operate on as they expand overseas. Anthropic chief executive Dario Amodei has been among the most vocal industry figures calling for binding safeguards, warning that without them the race between labs could outpace the ability of governments to respond. The company has published policy proposals and Amodei has testified before US senators on the risks posed by increasingly capable models.

Washington's answer, as described by the BBC, was no. The administration has favored a lighter approach that leaves room for American firms to keep their lead rather than signing onto frameworks negotiated with other capitals. That position puts the US at odds with the European Union, whose AI Act is being phased in, and with a growing list of countries that have drafted or adopted their own rules. It also deepens the split between the labs that want harmonized standards and a government that sees regulation as a drag on domestic competitiveness.

The practical effect for OpenAI and Anthropic is that any global standards they help shape will lack US government endorsement, which weakens their leverage when they sit down with counterparts in Europe, China or elsewhere. Both companies have said they support some form of international coordination on frontier models, including safety testing and reporting requirements, but their influence rests largely on voluntary commitments rather than treaty obligations. Rivals that oppose strict rules can now point to Washington's stance as evidence that binding global standards are not inevitable.

The decision also hands ammunition to critics who say the US is ceding the rules-writing role to Brussels and Beijing. American lawmakers have introduced a series of AI bills, but none has cleared Congress, and the vacuum has left states like California to act on their own. For the labs, the question now is whether they keep pressing for international alignment without their own government behind them, or shift to defending the voluntary approach Washington prefers.

What to watch is whether the administration's position holds as the EU enforces its AI Act and other countries finalize their own regimes. If American firms find themselves shut out of markets that demand compliance with foreign standards, the argument for US engagement could revive. Amodei and his peers have made clear they intend to keep making the case, even after this setback.