Colorado Chatbot Rulemaking Timeline: Disclosure and High-Risk AI Deadlines
Colorado runs two converging regulatory clocks for AI. First, the Colorado AI Act (SB 24-205) tasks the Attorney General with adopting rules governing high-risk AI systems, including the consumer-disclosure and algorithmic-discrimination duties that take effect in 2026. Second, Colorado's standalone chatbot-disclosure law (C.R.S. 6-1-716) requires that an automated system interacting with a consumer clearly disclose that it is not a human -- a rulemaking and enforcement track that now overlaps the AI Act's transparency regime. For a verification-first operator, the timeline matters because disclosure alone is not measurement.
A chatbot that announces 'I am an AI' has made a claim; it has not produced a verified measurement credential proving its boundary is instrumented, its refusal behavior is measured, and its incidents are logged. The Council of AI positions the Colorado rulemaking window as an answer-first procurement moment: deployers who arrive at the H2 2026 rulemaking hearings already carrying signed, per-axis measurement evidence (care, governance, refusal, sandbox-escape) satisfy the disclosure duty with auditable records instead of boilerplate notices. Key dates to watch: the Attorney General's high-risk rules are expected to finalize before year-end 2026, with enforcement commencing on a rolling basis thereafter; chatbot-disclosure complaints are already being resolved through the existing consumer-protection docket.
Standards and sources referenced
- Colorado Artificial Intelligence Act, SB 24-205 (2024)
- C.R.S. 6-1-716 -- chatbot and automated-system disclosure
- Council of AI Containment Incident Index (CONTAINMENT_INCIDENT_INDEX.json)
- CSOAI Verified Measurement Credential standard (DEFONEOS-SEAL)