AI Disclosure
The duty to tell consumers that an AI system played a material part in a decision that affects them.
AI disclosure is the duty to tell consumers that an AI system played a material part in a decision that affects them. Colorado SB 26-189 will carry that duty from January 1, 2027, and it attaches to automated decision-making technology, or ADMT, rather than to the “high-risk AI systems” named by the statute it repealed. What puts a system in scope is the decision it feeds, not the technique inside it, so a fixed scoring table can qualify while a far more sophisticated model used somewhere inconsequential does not.
Disclosure obligations are not uniform. Some laws require only that AI was used. Others require an explanation of the decision, the data considered, or the right to appeal. In insurance, the challenge is to provide meaningful notice without overwhelming consumers with technical detail.
AI disclosure is closely tied to consumer recourse. A consumer cannot contest a decision they do not know was AI-influenced. A Colorado carrier owes less of this than the headline suggests. Section 6-1-1708(1)(a) hands the practice of insurance back to the SB 21-169 regime, so the ADMT notice duties bite hardest on a carrier’s employment systems and on vendors deploying in their own right. Where they do apply, notices should be clear, timely, and aligned with whatever review process the consumer is being pointed at. See our glossary entries on Colorado SB 26-189, consumer recourse, and automated decision systems.