Colorado Moves to Scale Back Landmark AI Law Amid Industry Pushback
Key Takeaways
- Colorado's AI Policy Work Group released recommended Colorado AI law amendments that would narrow the scope of SB 205 by focusing on automated decision-making technology that "materially influences" consequential decisions rather than regulating all high-risk AI systems.
- The proposal removes the controversial algorithmic discrimination standard that industry groups criticized as too vague, instead relying on existing anti-discrimination laws to address discrimination arising from AI-driven decisions.
- Colorado SB 205 changes include scaled-back transparency requirements for developers and deployers, broader exemptions for routine business software, and a shift from risk-based regulation to a use-based model centered on consumer rights.
- The working group's recommendations on automated decision-making technology regulation maintain consumer disclosure requirements but simplify them, requiring clear disclosures at the point of interaction rather than detailed upfront notices.
- Lawmakers have until May 13 to reach consensus on amendments before the law's June 30 effective date, with federal preemption efforts adding pressure to demonstrate that state AI policy working group solutions can succeed.
- If you're a subscriber, click here for the full edition of this update. Or, click here to learn more about our MultiState.ai+ subscription.
Colorado was the first state to enact a comprehensive regulatory framework for artificial intelligence in 2024, but nearly two years later, SB 205 has yet to take effect. Lawmakers have acknowledged the need for changes, but have struggled to agree on how to maintain meaningful guardrails while taking a lighter-touch approach to a rapidly evolving industry. Last fall, Governor Jared Polis (D) organized a working group to address issues with the law, and last week the group released recommended legislation. The proposal would scale back some of the requirements that industry groups had criticized, though it remains unclear whether it will provide a path forward for lawmakers.
How we got here
Back in 2024, several states pursued legislation targeting algorithmic bias in AI systems, introducing the concepts of "high-risk" AI and "consequential decisions" into the regulatory debate. While industry attention focused on stopping Connecticut's proposal (CT SB 2), Colorado lawmakers moved quickly on a similar bill modeled after the Connecticut framework. Both chambers approved an amended version of CO SB 205 in the final two days of session, and Governor Polis signed it into law — albeit with reservations — making Colorado the first state to enact a comprehensive AI regulatory framework.
If you're a subscriber, click here for the full edition of this update. Or, click here to learn more about our MultiState.ai+ subscription.
