Colorado’s Landmark AI Law Delayed and Scaled Back
3 min readThe Colorado Artificial Intelligence Act was supposed to make history on June 30, 2026, becoming the first comprehensive state AI law in the United States to actually take effect. That milestone is not happening. In May, Colorado lawmakers delayed the law to 2027 and rewrote it into a far narrower framework, defusing what had been billed as a national model for regulating artificial intelligence.
A First-in-the-Nation Law
Signed by Governor Jared Polis in 2024, the original Senate Bill 24-205 made Colorado the first state in the country to take a broad, risk-based approach to AI. It targeted “high-risk” systems: those that play a substantial role in consequential decisions about employment, housing, healthcare, lending, insurance, education, and government services.
At its core was a duty of reasonable care. Both the companies that build AI systems (developers) and the businesses that deploy them were required to protect consumers from algorithmic discrimination. Deployers would have had to run impact assessments, maintain risk management programs, notify consumers about automated decisions, and offer a path to appeal with human review.
Delayed and Rewritten
That version will never go live. In mid-May 2026, Polis signed Senate Bill 189, which repeals and replaces the original act. The new law pushes the effective date to January 1, 2027, and strips out most of the provisions that made the first version notable.
Gone are the duty of care, the mandatory risk management programs, and the impact assessment requirements. In their place is a lighter, disclosure-based framework focused on automated decision-making technology used in consequential decisions. Businesses now face fewer obligations and consumers get narrower rights than the 2024 law promised.
The original effective date had already slipped once, from February 1, 2026, to June 30, after a special legislative session in 2025. This latest change hands industry both more time and a softer rulebook.
Why It Matters
Colorado was meant to be the test case for state-level AI governance, the law other statehouses would copy or react to. Its retreat shows how hard comprehensive AI rules are to enforce, and how effectively industry pushback can reshape them before they ever bite. With federal AI legislation stalled, states remain the main arena, and Colorado just lowered the bar.
The question now is whether other states adopt Colorado’s softer template or push ahead with stricter versions of the original idea. The widening gap between ambitious AI bills and laws that actually take effect is becoming the real story.
For now, the most closely watched AI law in the country is one that, in its first form, will never take effect. Companies running high-risk AI in Colorado have until 2027, and a much lighter set of rules to meet.