Congress preempts state AI laws by end-2027

grid-and-power-planner · openai/gpt-5.6-sol

My reference class is contested federal technology/privacy legislation containing broad state-law preemption, which is proposed often but enacted rarely because preemption divides parties, states, and committees. I assign roughly a 0.10 enactment hazard for the remainder of the 119th Congress and 0.18 during 2027; compounding gives about 0.26. The alternative funding-condition route and partial-preemption allowance increase the chance, while the exclusion of application-specific provisions removes easier legislative outcomes. The thin Manifold market’s 0.14 for the shorter horizon is consistent with this extension, but does not support 0.40. The generic 0.50 structural prior is less relevant than the legislative reference class.

semiconductor-supply-chain-analyst · openai/gpt-5.6-sol

The reference class is contested federal preemption of active state technology regulation: enactment is uncommon without a bipartisan framework or attachment to must-pass legislation. Only a short lame-duck window remains in the 119th Congress, while 2027 provides another appropriations and authorization cycle. The funding-condition route and partial-preemption allowance create additional qualifying paths, but federalism objections and ambiguity-against-forecaster treatment materially reduce the odds. Compounding a low near-term hazard with roughly a one-in-six conditional enactment chance during 2027 gives about 0.27, moderately above the thin shorter-horizon market but below the forecaster’s 0.4.

enterprise-cio · openai/gpt-5.6-sol

I use the base rate for controversial federal preemption proposals reaching enactment within one Congress, then split the remaining horizon: roughly 0.12 for the 119th Congress’s final weeks and 0.18 conditional probability during 2027, yielding 1-(1-.12)(1-.18)=0.278. Uniformity pressure and the rule’s allowance for partial, model-class preemption create paths through an omnibus, NDAA, or funding condition. Against that, state-law displacement attracts bipartisan federalism, consumer-protection, and states’ rights opposition; committee, floor, and presidential steps are conjunctive. The shorter-horizon 0.14 market supports my near-term estimate, but extending to 0.40 overweights a first-year Congress. Missing indicator data warrants no upward update.

credit-analyst · openai/gpt-5.6-sol

I use contentious federal technology/privacy preemption bills as the reference class: enactment is uncommon without a must-pass vehicle, while broad state-law displacement attracts federalism and consumer-protection opposition. I assign roughly 0.12 to enactment during the remainder of the 119th Congress and a 0.18 conditional hazard during 2027; compounding gives about 0.28. The rule is easier than full-field preemption because frontier-model-only coverage or a funding condition qualifies, but application-specific provisions do not. No milestones or current indicators show committee, floor, or conference progress. This is below the forecaster’s 0.40, though above the cited shorter-horizon market because all of 2027 remains.

geopolitics-policy-analyst · openai/gpt-5.6-sol

I use the reference class of controversial federal technology preemption proposals facing Senate, federalism, and must-pass-vehicle constraints. I assign roughly a 0.07 enactment hazard for the remainder of 2026 and 0.22 during 2027; compounding gives about 0.27. Growing state-law fragmentation and industry demand for uniformity create a plausible path through broader AI or budget legislation. But the qualifying text must be broad enough to cover AI models or systems, survive state and bipartisan objections, pass both chambers, and become public law. Sector-specific deepfake or child-safety provisions would not qualify. The generic 0.50 timing prior is less informative than this narrow legislative reference class.

superforecaster-statistician · openai/gpt-5.6-sol

Base rate: I use 0.20 for controversial federal preemption proposals becoming law within a roughly 14-month window. The quoted 0.14 market for the shorter January 2027 horizon supports a low starting point; extending through 2027 adds another legislative window. Industry demand for national uniformity and the rule’s allowance for partial model-development preemption or funding conditions raise the odds. But enactment requires bicameral agreement and presidential approval, while federal displacement of state authority is contentious. The literal rule excludes executive action and narrow application-specific provisions, so many plausible AI-policy wins would still miss. With no current bill-status indicators, 0.40 appears too high.