White House Denies Approving AI Model Releases, Says Decisions Rest With Companies

A White House official told reporters that the administration does not formally approve AI model releases from private companies, insisting that any government engagement, testing or coordination remains voluntary. According to the official, final decisions on timing and scope of AI releases rest entirely with the companies themselves, pointing to a recent executive order from President Trump as the governing framework.

Recent Actions Tell a More Complicated Story

Despite those assurances, the administration blocked access to two advanced AI models last month over national security concerns, only restoring availability after weeks of negotiation with the developer. Separately, another major AI company said it would limit access to new models to a group of trusted partners in order to comply with government requests, suggesting government influence over release decisions remains significant in practice.

Competitive Pressure From China Intensifies

The debate comes as Chinese AI developers close the performance gap with leading American labs. A newly released open-weight model from a Beijing-based startup matched or outperformed top U.S. systems on several benchmarks this week, prompting concern among some former administration advisors that overly cautious U.S. policy could cede ground in the global AI race.

New Cybersecurity Initiative Raises Oversight Questions

This week, the administration launched a new program focused on identifying and fixing cybersecurity vulnerabilities in collaboration with private AI companies. According to a person familiar with the matter, the initiative would effectively give the White House authority over which companies can access newly developed AI models, raising questions about the future of existing industry-led safety collaboration efforts.