US lawmakers are watching to see whether Colorado's bill for regulating AI, modeled on Connecticut's now derailed bill, can withstand the tech lobby pressure
Mallory Culhane / Politico : X: @cendemtech , @mallorykculhane , and @cendemtech . LinkedIn: James Maroney X: @cendemtech : @CenDemTech's Matt Scherer in @Politico talking about recent state legislative efforts in Colorado and Connecticut to prohibit discrimination in the workplace caused by #AI. https://www.politico.com/... Mallory Culhane / @mallorykculhane : spent a weird amount of time thinking about what Connecticut and Colorado are up to on AI for the past few months. and you can (should) read all about how the two states are leading the charge to build momentum for a national model: https://www.politico.com/... @cendemtech : @politico Matt Scherer: “The bill is not perfect in Colorado by any stretch of the imagination. I still think that, for example, its enforcement provisions are too weak ... but I think that the Colorado bill provides enough to be a good starting point.” https://www.politico.com/... LinkedIn: James Maroney : This past year I worked on SB 2, a comprehensive AI regulatory framework. I am disappointed that we were unable to get a vote in the house …
Context & Ripple Effects
Colorado and Connecticut are being positioned as state-level test beds for an AI anti-discrimination framework after Connecticut's SB 2 failed to reach a House vote. That makes Colorado's legislative durability, rather than the mere introduction of another bill, the key signal for policymakers seeking a national model.
The fight also follows reporting that proposed state AI-bias measures contained accountability loopholes that could limit enforcement. With federal lawmakers still struggling to engage deeply with AI governance in earlier coverage, state legislative experiments carry added weight.
First-order effects
- Colorado lawmakers and the bill's supporters face an immediate test of whether a discrimination-focused AI framework can survive organized tech-industry opposition.
- Connecticut's stalled SB 2 becomes a cautionary reference point: weak enforcement may make a proposal more politically viable, but can also narrow its practical impact.
Second-order effects
- Other state policymakers considering AI-bias rules can use Colorado's outcome to judge which obligations and enforcement provisions attract the strongest lobbying resistance.
- Tech lobby groups gain an incentive to press for narrower duties, exemptions, or softer enforcement in state proposals, especially where lawmakers want a model that can be replicated nationally.
Third-order effects
- If state bills repeatedly stall or emerge with limited enforcement, the US could develop a fragmented AI-governance landscape in which protections vary substantially by state rather than converging on a durable national baseline.
- Conversely, a workable Colorado model could give federal lawmakers a concrete state template; whether it does so depends on its final enforceability, not simply its passage.
The trend: AI governance is moving through state-level experimentation, with anti-discrimination rules becoming an early contest over whether enforceable safeguards can withstand industry lobbying.