AI Policy Newsletter - March 26, 2026

• On March 20, The White House released its National AI Legislative Framework, which entails recommendations to Congress including online child safeguards, streamlined data center permitting for on-site power generation, anti-censorship provisions, intellectual property protections, and AI workforce training expansion. The framework urges Congress to preempt state AI laws that "impose undue burdens" while preserving state authority over traditional police powers, zoning, and procurement. Notably, it writes that “States should not be permitted to regulate AI development, because it is an inherently interstate phenomenon with key foreign policy and national security implications.” In an interview with Bloomberg on Thursday morning, David Sacks suggested that Congress may be able to pass bipartisan legislation creating standards for AI in the next serval months, noting a positive response from Congress on the White House framework.   

  • Some key provisions mirror Sen. Blackburn's TRUMP AMERICA AI Act discussion draft released on March 18. Among other things, the TRUMP AMERICA AI Act would sunset Section 230, create new AI developer liability pathways, exclude unauthorized copyrighted works from fair use in AI training, mandate third-party audits to prevent political bias, restrict federal procurement of ideologically biased models, and direct the Energy Secretary to enter agreements with data center operators protecting consumers from rate increases. The framework additionally incorporates the Kids Online Safety Act including duty of care language omitted by the House KOSA bill, and the NO FAKES Act.

  • It is important to note that both the framework and the Blackburn draft would need to be legislated by Congress and would require Democratic support in the Senate to advance.

• On March 16, The Department of Commerce announced a new phase of the American AI Exports Program, calling for proposals for full-stack AI technology packages. The call for proposals includes two types of industry-led consortia: pre-set consortium demonstrating capability across all layers of the AI technology stack to maintain global offerings ready for deployment on an ongoing basis and on-demand consortium formed by industry as “custom-made” options for specific opportunities identified by the Program. The Department will begin accepting proposals on April 1 and accept them for 90 days thereafter.

Download the full newsletter here.