The Trump administration is actively discussing government oversight of artificial intelligence models before they are released to the public, according to a New York Times report published Sunday citing U.S. officials and people briefed on the deliberations. The discussions include a possible executive order to establish a working group on AI that would coordinate review processes across federal agencies.

Reuters, Bloomberg, and Forbes all confirmed the report within hours, signaling this is not a trial balloon but a genuine policy discussion at the highest levels of the White House.

The 180 Nobody Saw Coming

This is a dramatic reversal. The Trump administration has spent its time in office positioning itself as the anti-regulation AI government. It revoked Biden-era AI safety executive orders. It championed American AI dominance as a national priority. It welcomed Big Tech CEOs to the White House and promised to unleash innovation.

Now the same administration is discussing pre-release vetting of AI models. The kind of oversight that would make the Biden White House blush.

The timing is impossible to ignore. This news drops while the Musk v. OpenAI trial plays out in Oakland, while Anthropic signs a $1.5 billion consulting joint venture, while the Pentagon inked AI deals with eight major tech companies last week, and while state legislatures from Maryland to Connecticut are racing to pass their own AI regulation. The federal government appears to have concluded that if it does not act, the states will create a patchwork of rules that could be worse for industry.

What the Executive Order Could Look Like

Details remain thin, but the contours are becoming clear. The executive order under discussion would create an interagency working group tasked with developing a framework for evaluating frontier AI models before public deployment. This is not the lightweight "voluntary commitments" approach from previous administrations. This is structural.

The key question is scope. Does "vetting" mean safety testing for catastrophic capabilities? Does it mean reviewing models for bias, misinformation, or election interference potential? Does it mean national security screening to prevent adversaries from exploiting open-weight models? The answer will determine whether this becomes a meaningful safeguard or bureaucratic theater.

Industry Reaction Will Be Swift

Expect the lobbying to start immediately. OpenAI, Google, Anthropic, and Meta have all positioned differently on regulation. Anthropic has long argued for government oversight of frontier models. OpenAI has been more ambivalent, supporting safety frameworks while fighting any rules that would slow its commercial rollout. Google has tried to occupy the middle ground. Meta, under its open-source strategy, has the most to lose from pre-release vetting requirements.

The open-source AI community will be watching closely. If pre-release vetting applies to open-weight models like Meta Llama, it could fundamentally change how those models reach developers and researchers worldwide.

What to Watch

No executive order has been signed yet. These are discussions, not decisions. But the fact that they are happening at all tells you something important: even the most pro-business, anti-regulation administration in modern history has looked at what AI companies are building and concluded that the government needs a seat at the table before these models ship.

That alone is the story.