U.S. enacts AI safety bill before 2027?

The strongest reason to take a Yes outcome seriously is that this is no longer a purely hypothetical AI policy discussion. There are active bipartisan negotiations, leadership figures are engaged, and the draft appears to be centered on real regulatory concepts such as duty of care, mitigation obligations, and possible court-based blocking mechanisms for unsafe model releases. That matters because a bill with any of the listed provisions would qualify, so the threshold for Yes is not a sweeping omnibus AI law but a narrower safety-focused statute that could emerge from the current talks if negotiators converge quickly enough.

The main problem for Yes is timing, and timing is doing a lot of work against the market. The bill is not finalized, key language is still unresolved, and multiple reports say the chance of moving quickly is very low. Even if the Senate gets closer to agreement, the House appears behind schedule and may not act until later in the year or even into lame-duck territory. In practice, a federal AI bill has to clear both chambers, survive procedural friction, and get signed before the end of 2026, which is an extremely compressed path for a first-of-its-kind policy effort.

Historical and political context also point toward caution. Congress has spent years discussing AI governance without enacting a comprehensive federal framework, while state-level regulation has grown and can reduce the urgency for a rushed federal deal. That does not make passage impossible, but it does mean the default legislative pattern has been delay rather than breakthrough. The current proposal may be the most credible federal safety effort so far, yet the market should discount it heavily because bipartisan interest is necessary but not sufficient, and the remaining disputes are exactly the kinds that often sink or postpone major bills.

Compared with the market price, I think the Yes probability should be modestly above zero but still clearly low. The current 12.5% implied by the market seems directionally reasonable, but I would assign a somewhat higher chance because the negotiations are real and the coalition appears more functional than in earlier AI policy pushes. Still, absent a fast breakthrough in the next legislative stretch, the most likely outcome remains that Congress either misses the deadline or passes something that does not reach the required form in time.