US Government Urges AI Companies to Self-Regulate Amidst Regulatory Uncertainty

The Trump administration has made it clear that it expects AI companies to take responsibility for their own safety measures rather than relying on federal regulation. In remarks made by Vice President JD Vance from September 16 to 18, 2026, he urged developers to avoid creating potentially dangerous technologies, coining the phrase "stop building Frankenstein" to encapsulate this sentiment.

President Trump has downplayed concerns about AI posing existential threats, labeling such warnings a "hoax" and instead promoting AI as a significant economic opportunity. Supporting this approach, a June 2, 2026 executive order established a voluntary framework for reviewing high-risk AI models, allowing developers to submit their systems for government evaluation with a streamlined vetting process, reduced from 90 days to a maximum of 30 days.

Some major AI firms, including OpenAI and Anthropic, have begun implementing their own internal safety protocols, such as appointing "embedded evaluators" to assess risks before deployment. However, the industry remains divided on the adequacy of self-regulation, with some congressional members advocating for independent oversight, although these proposals face significant challenges.

The administration's preference for collaboration over stringent regulation is influenced by the ongoing technological competition with China, as imposing heavy compliance requirements could disadvantage US companies.