What happened
On 6 October 2026 the Colorado Department of Law (Attorney General) released an interim redline of its proposed rules implementing the Automated Decision-Making Technology in Consequential Decisions Act (SB 26-189) and the Conversational AI Services (Chatbot Safety) Act (HB 26-1263), following 60+ comments on the 11 August draft. The redline redefines 'de minimis factor', adds a 'materially influence' presumption, and revises multiparty-arrangement rules, retaining strict treatment of 'financial or lending service' decisions. The formal rulemaking hearing is 26 October 2026 (comments due that day); the AG must finalise rules by 1 January 2027 and both laws take effect 1 January 2027.
Why it matters
Colorado is the leading US state implementing binding AI decisional-oversight law: the ADMT Act imposes risk-assessment, impact-assessment and disclosure obligations for AI in consequential decisions (credit, employment, housing), and the Chatbot Safety Act imposes duties on AI chatbots that materially influence consumer decisions. Financial-services firms and any deployer of AI in consequential decisions face an imminent compliance build-out regardless of the federal picture.
Action needed
Substantive comments on the interim redline are due 11:59 p.m. MST on 26 October 2026; entities subject to ADMT/Chatbot obligations should comment now and begin building the assessment/disclosure/curation workflows before the 1 January 2027 effective date.