In a new letter, Senator Elizabeth Warren accuses major AI firms of pressuring the Trump administration to weaken AI transparency rules within the USMCA framework, arguing regulators should have broader access to information about AI models.

The appeal comes amid rising calls for tougher AI regulation after recent incidents and lawsuits involving AI systems.

Her argument cites a CSIS analysis of a pattern in recent US bilateral trade deals, describing a 'containment doctrine' that could undermine domestic tech regulations like the EU's Digital Markets Act and Digital Services Act.

Warren argues that removing secrecy around source code is essential so governments can better protect citizens from AI-related harms as AI activity crosses borders.

Industry voices counter by framing DMA-like measures as protectionist and warn such rules could invite trade challenges, while the administration signals willingness to use leverage, including potential tariffs on digital services taxes.

The USTR has not commented on Warren’s letter, and lawmakers and industry groups remain divided over what information should be disclosed for AI oversight.

(Credit line: Reuters)

No immediate comment was obtained from the USTR.

The USTR has not commented on Warren’s letter at this time.

Industry groups, led by the Computer & Communications Industry Association and the Telecommunications Industry Association, argue disclosure is necessary for reliable and safe global AI services while cautioning about potential impacts on the AI supply chain.

Reuters reports Warren’s accusation, but the specific firms and USTR have not publicly addressed the claim, leaving the outcome uncertain.

CSIS notes that the approach has limited impact on large economies but has penetrated smaller ones, with interpretation and effectiveness varying by country and market size.