Will the U.S. federal government enact legislation that substantively regulates or authorizes regulation of artificial intelligence systems or their development before 1 January 2026?

The current U.S. federal AI regulatory landscape lacks comprehensive legislation, with governance fragmented across executive actions, targeted bills, and state-level initiatives (White & Case). Consequently, federal agencies have relied on existing laws and voluntary frameworks to implement any AI-specific rules (SIG). Congress has introduced AI-related bills like the TAKE IT DOWN Act (prohibiting the publishing of nonconsensual AI-generated intimate imagery, now passed by the Senate and the House) and the CREATE AI Act (codifying a national AI research resource), but these specific bills address the publication of AI-generated content and establish resources to facilitate development, and consequently, they do not directly regulate AI systems or their development. Broader regulatory frameworks for AI systems have yet to be established (Inside Global Tech, CBS News).

Resolution Criteria: This question will be resolved as “Yes” if, before 1 January 2026, both chambers of the U.S. Congress pass a bill substantively regulating AI development or restricting its use AND the President signs it into law or Congress overrides the President’s veto.
To count for resolution, a bill must contain provisions that explicitly regulate artificial intelligence systems, models, or development. To be considered as “substantively regulating artificial intelligence systems or their development” qualifying provisions must:

• Explicitly mention AI terminology, such as “artificial intelligence,” “AI,” ”generative AI,” “foundation models,” “machine learning,” or other clearly identifiable AI-specific terminology
• Create binding requirements, for example, by establishing one or more of the following:
  • Mandatory standards or thresholds for AI system performance, safety, or transparency
  • Prohibited uses or applications of technologies
  • Compliance mechanisms with enforcement authority
• Go beyond general technology regulations and specifically target AI capabilities, processes, or risks.

The following will not be sufficient for resolution: 

• Executive orders, new agency regulations implementing existing law, and non-binding guidelines regarding the development of AI
• General technology legislation (e.g., new data privacy or cybersecurity laws) that affects AI systems, but does not regulate them specifically
• Provisions allocating funding for AI research and development or establishing committees to study AI without any regulatory components
• Provisions that simply define AI

In situations not fully addressed above, the RFI team will determine the resolution based on whether the legislation aligns with the spirit of "substantive regulation” of AI systems or their development, as outlined in these criteria.