Trump Administration Shifts Course on AI Regulation
The White House is considering tighter government oversight of artificial intelligence models, marking a reversal from the Trump administration's initial hands-off approach to tech companies.
In his first month in office, President Trump revoked a 2023 Biden-era executive order designed to protect Americans from AI risks. The administration called the order an obstacle to innovation. But now, according to reporting by the New York Times, the White House is weighing an executive order that would increase oversight of AI models before their public release, including a formal government review process.
A White House official told MarketWatch that discussion of such an order is "speculation," and any announcement will come from the president.
Voluntary Agreements Take Shape
The administration is moving forward with a less formal approach: voluntary agreements with major AI developers. On Tuesday, the Center for AI Standards and Innovation within the Department of Commerce announced partnerships with Google DeepMind, Microsoft, and xAI. These companies will submit models for "pre-deployment evaluations" before public release.
The shift reflects growing public skepticism about AI and mounting concerns over job losses and energy-intensive data centers. National security considerations also factor in.
Why the Change?
Darrell West, a senior fellow at the Brookings Institution, called the pivot "a startling turn from his libertarian stance of the past year." The departure of David Sacks, Trump's top AI adviser and a deregulation advocate, in March may have opened space for more regulatory thinking.
The immediate trigger appears to be national security concerns. James Lewis, a senior adviser at the Center for Strategic and International Studies, said the administration's alarm over Anthropic's release of a model called Mythos - which can discover and exploit cybersecurity vulnerabilities - likely drove the policy shift. Anthropic restricted public access but shared the model with select companies through Project Glasswing.
"I think a lot of it comes from Anthropic persuading the Defense Department that this was a risk that needed to be seriously considered," Lewis said.
Limits to Oversight
Lewis expects the Trump administration will focus narrowly on national security rather than adopt the broader Biden approach, which included safety, privacy, and equity concerns. "This administration is a little allergic to government surveillance," he said, so oversight will likely stop short of intrusive monitoring.
The CAISI partnership remains voluntary and non-regulatory. Companies cannot be forced to participate, and Lewis suggested the program might offer legal protections similar to those in the Cybersecurity Information Sharing Act, encouraging voluntary disclosure.
Industry Pushback
Tech advocates worry the shift signals a larger regulatory trend. Daniel Castro, president of the Information Technology and Innovation Foundation, said pre-approval requirements could slow U.S. development and give competing nations an advantage.
"It effectively forces firms to seek permission to innovate, slowing development to the pace of Washington rather than the market," Castro said. "Other countries will keep advancing, and open-weight models released abroad won't face similar constraints."
For managers overseeing AI initiatives, the policy uncertainty creates immediate planning questions. Understanding how regulatory shifts affect AI strategy has become essential for business continuity. Technology leaders should also review governance frameworks and compliance requirements as oversight requirements evolve.
Your membership also unlocks: