Written comments on Colorado’s proposed ADMT and Chatbot Safety rules must land by 11:59 p.m. MT tonight to be guaranteed presentation at the October 26 rulemaking hearing. After that, the record closes to structured input, and whatever the Attorney General finalizes before January 1, 2027 is what governs lead scoring, screening tools, and customer-facing chatbots across the state.

The rules, filed August 11 by the Colorado Department of Law, implement SB 26-189 and HB 26-1263. They’re the operational layer sitting on top of the revised ADMT and chatbot rules filed in August, and they cover any AI that “materially influences” a consequential decision. The AG is openly asking the public to pick between two competing definitions of that phrase. The lower threshold pulls in most AI-assisted employment, credit, and housing uses. The higher “substantial factor” test is narrower. Whichever wins today’s comment cycle shapes everything downstream.

For a 5–30 person operator, the compliance math is heavier than the statutes read. Adverse-outcome notices must go out within 30 days through two separate channels. Consumer rights to data access, correction, and human review must be completed within 45 days. Chatbot sessions require AI disclosure at the start of each day’s first interaction, every three hours in continuous sessions, and whenever a user asks. Narrow menu- and rule-based bots can escape under specific conditions; conversational ones can’t.

Seyfarth’s read is that the rules demand “significantly more compliance infrastructure than the statutes alone suggest”, decision-level explainability, staffed human review, vendor documentation pass-through. The hearing itself runs October 26 at 10:00 a.m., with comments still accepted through 11:59 p.m. MT that night. But presentation rights expire tonight. Rulemaking records are shaped by who shows up on the deadline, not by who objects after.

Sources

Sources