Will both chambers of the 119th Congress pass legislation containing a qualifying frontier-AI safety provision between September 29, 2026 and January 3, 2027?

probability: 4%

rationale:

The estimated probability is 4%. Initial legislative parameters and current Congressional dynamics dictate a very low baseline for passing a qualifying frontier-AI safety provision before January 3, 2027, consistent with the constrained lame-duck legislative window. Both chambers must pass binding, generally applicable text targeting catastrophic risk within that window congress.gov. While legislative vehicles like the FRONTIER Act (H.R. 9925) and an unreleased Senate draft led by Senators Thune, Cruz, and Klobuchar possess qualifying language—such as incident reporting, third-party audits, and civil liability—procedural advancement remains thoroughly stalled 2 sources.

The established context shows nearly insurmountable political headwinds for standalone passage. Executive and House leadership exhibit overt hostility toward binding federal AI mandates. Former President Trump has labeled AI catastrophe fears a "hoax" and favors industry self-regulation, a position mirrored by House Speaker Johnson and Senate Majority Whip Barrasso 2 sources. Consequently, major, novel regulatory frameworks are highly unlikely to bypass standard committee processes to reach the floor without explicit bipartisan leadership consensus, which is currently absent.

Must-Pass Vehicles and Preemption The primary path to enactment rests entirely on an NDAA conference compromise or an omnibus spending package. Existing FY2027 NDAA AI provisions are strictly DoD- or procurement-focused, which fails the general applicability test law-ai.org. The sole viable maneuver involves a negotiated lame-duck deal pairing a federal catastrophic-risk standard with the preemption of state AI laws, an approach currently favored by industry leaders seeking regulatory uniformity punchbowl.news. However, ongoing disputes over liability thresholds and whether government safety testing should be mandatory or voluntary 2 sources drastically limit the viability of this route.

Ultimately, the conjunctive burden of both chambers passing qualifying text within weeks is structurally prohibitive. Even if preemption forces a lame-duck compromise, the risk of the resulting standard defaulting to a non-qualifying voluntary framework is extremely high 2 sources. The structural, temporal, and political barriers, viewed consistently with identical forecasts assessing this constrained lame-duck legislative window, firmly cap the likelihood of success at 4%.