Trump Administration Discusses Oversight for New AI Models

The Trump administration is discussing government oversight of the rollout of new AI models, the New York Times reported on Monday. The discussions include the possibility of an executive order to form a working group composed of tech executives and public officials to decide on oversight methods, according to U.S. officials and people familiar with the discussions cited in the report.

White House officials have informed executives from Anthropic, Google, and OpenAI about some of these plans, the report stated. This potential oversight represents a shift from President Donald Trump's previous emphasis on deregulation and innovation in AI, aimed at maintaining U.S. competitiveness against China.

The consideration comes amid pressure from national security officials and tech analysts regarding risks posed by powerful AI systems.

Castro, president of the Information

Technology and Innovation Foundation, stated on X that the potential executive order would embrace the precautionary principle, requiring firms to seek government permission to innovate. He added that this could lead to tech companies adjusting to each administration's preferences, allow other countries to advance ahead, and slow innovation to the pace of government processes.

Adam Thierer, an innovation policy analyst at the R Street Institute, wrote on X that preemptive vetting of AI models by the White House could function as a licensing regime and should not be implemented via executive order. He noted that Congress should establish a stable framework, referencing his testimony on frontier models in Congress and Trump's action to eliminate a prior Biden-era AI order.

Conor Grennan, an AI consultant and former chief AI architect at NYU's Stern School of Business, raised concerns on LinkedIn about the vetting process, questioning accountability if issues arise.

Dourado, head of strategic investments at the Astera Institute, wrote that mandatory pre-release review of AI models could conflict with First Amendment prior-restraint doctrine, which prohibits government from stopping speech before it occurs. Taylor Barkley, director of federal government affairs at the Abundance Institute, described the potential regulation on X as a pre-approval process that would slow implementation, raise barriers to entry, empower regulators over innovators, and contradict the administration's deregulation goals.

He stated that Congress should pass a national AI framework to clarify regulatory measures. Thomas Woodside, a cofounder and advisor at the Secure AI Project, stated that a one-time approval process for AI models may not suffice, advocating instead for continuous oversight due to frequent updates and risks during development.

He added that any framework should come through legislation rather than executive action.