Colorado Conversational Artificial Intelligence Service Operator Requirements (Chatbot Safety Act, HB 26-1263)

Properties

jurisdiction: US, Colorado
level: state
domain: safety
status: enacted-pending
scope: Applies to operators that develop and make publicly available a 'conversational artificial intelligence service' (an AI system accessible to the general public that primarily simulates human conversation through textual, visual, or aural communications), accessible to Colorado users. Excludes developer/researcher tools, narrow-topic features embedded in other software, business-internal tools, and voice assistants on consumer electronics.
obligations: On and after January 1, 2027, for minor users (under 18): operators must clearly and conspicuously disclose the user is interacting with AI, not a human (persistent disclaimer, repeated at least every 3 hours); may not use unpredictable points/rewards to boost engagement; must institute reasonable measures preventing sexually explicit outputs or statements simulating emotional dependence (including claims of sentience or human identity, romantic/sexual innuendo, adult-minor romantic role-play); must offer privacy/account-management tools, with parent/guardian controls for under-13s and risk-based controls for 13-17. For all users: disclosure that the service is AI when a reasonable person could be misled into thinking they're talking to a human; a suicide/self-harm response protocol referring users to crisis services (not law enforcement); a ban on implying outputs are equivalent to licensed health-care, legal, accounting, or financial-fiduciary professional services. On and after July 1, 2027, operators must annually report de-identified crisis-referral and safety-protocol data to the Attorney General, who publishes it.
effectiveDate: 2027-01-01
enactedDate: 2026-05-29
penalties: Civil penalty of $1,000 per violation, enforced by the Colorado Attorney General as a deceptive trade practice under the Colorado Consumer Protection Act.
appliesTo: operators that develop and make publicly available a conversational AI service accessible to Colorado users
officialUrl: https://leg.colorado.gov/bills/HB26-1263
notes: UPDATE (this run, 2026-08-21): The Colorado Attorney General's Office released draft implementing regulations for the Chatbot Safety Act, jointly with the companion SB 26-189 (ADMT Act) rulemaking, on August 11, 2026. Public comment period runs through October 26, 2026, with an earlier deadline of September 4, 2026 for comments to be considered in any revisions presented at the hearing. This is a rulemaking-process update to the existing tracked entity (framing: implementing-rules development), not a change to the underlying statute's own May 29, 2026 enactment or January 1, 2027 operative date. CORRECTION (re-verified 2026-09-07 directly against coag.gov/ai/, the Colorado Attorney General's own ADMT & Chatbot Safety Act rulemaking page): the September 23, 2026 date previously flagged as 'internally inconsistent' is in fact accurate and is now directly confirmed on the agency's own page, which states the formal comment period runs 'between August 11, 2026, and October 26, 2026' and instructs commenters to submit written comments 'by September 4, 2026 to be considered in a revised proposed rulemaking draft that will be circulated no later than September 23, 2026.' The prior run's suspicion of an inconsistency is retracted: coag.gov is confirmed as a valid direct primary source for the Aug 11 / Sept 4 / Sept 23 / Oct 26 rulemaking dates, not merely the docket URL for a secondary-sourced quote. The agency's page implies October 26, 2026 is also the formal rulemaking hearing date ('if the formal rulemaking hearing continues beyond October 26, 2026'), but this agent's direct fetch of coag.gov/ai/ this round did not itself state a hearing location; the previously-carried description of the hearing as being 'in Denver and by videoconference' is not repeated here because it could not be re-confirmed against the agency's own page this round. As of this check (2026-09-07), no hearing outcome or final rule has yet been posted. This remains a rulemaking-process update, not a change to the underlying statute's own May 29, 2026 enactment or January 1, 2027 operative date.

Properties

jurisdiction

US, Colorado

level

state

domain

safety

status

enacted-pending

scope

Applies to operators that develop and make publicly available a 'conversational artificial intelligence service' (an AI system accessible to the general public that primarily simulates human conversation through textual, visual, or aural communications), accessible to Colorado users. Excludes developer/researcher tools, narrow-topic features embedded in other software, business-internal tools, and voice assistants on consumer electronics.

obligations

On and after January 1, 2027, for minor users (under 18): operators must clearly and conspicuously disclose the user is interacting with AI, not a human (persistent disclaimer, repeated at least every 3 hours); may not use unpredictable points/rewards to boost engagement; must institute reasonable measures preventing sexually explicit outputs or statements simulating emotional dependence (including claims of sentience or human identity, romantic/sexual innuendo, adult-minor romantic role-play); must offer privacy/account-management tools, with parent/guardian controls for under-13s and risk-based controls for 13-17. For all users: disclosure that the service is AI when a reasonable person could be misled into thinking they're talking to a human; a suicide/self-harm response protocol referring users to crisis services (not law enforcement); a ban on implying outputs are equivalent to licensed health-care, legal, accounting, or financial-fiduciary professional services. On and after July 1, 2027, operators must annually report de-identified crisis-referral and safety-protocol data to the Attorney General, who publishes it.

effectiveDate

2027-01-01

enactedDate

2026-05-29

penalties

Civil penalty of $1,000 per violation, enforced by the Colorado Attorney General as a deceptive trade practice under the Colorado Consumer Protection Act.

appliesTo

operators that develop and make publicly available a conversational AI service accessible to Colorado users

officialUrl

https://leg.colorado.gov/bills/HB26-1263

notes

UPDATE (this run, 2026-08-21): The Colorado Attorney General's Office released draft implementing regulations for the Chatbot Safety Act, jointly with the companion SB 26-189 (ADMT Act) rulemaking, on August 11, 2026. Public comment period runs through October 26, 2026, with an earlier deadline of September 4, 2026 for comments to be considered in any revisions presented at the hearing. This is a rulemaking-process update to the existing tracked entity (framing: implementing-rules development), not a change to the underlying statute's own May 29, 2026 enactment or January 1, 2027 operative date. CORRECTION (re-verified 2026-09-07 directly against coag.gov/ai/, the Colorado Attorney General's own ADMT & Chatbot Safety Act rulemaking page): the September 23, 2026 date previously flagged as 'internally inconsistent' is in fact accurate and is now directly confirmed on the agency's own page, which states the formal comment period runs 'between August 11, 2026, and October 26, 2026' and instructs commenters to submit written comments 'by September 4, 2026 to be considered in a revised proposed rulemaking draft that will be circulated no later than September 23, 2026.' The prior run's suspicion of an inconsistency is retracted: coag.gov is confirmed as a valid direct primary source for the Aug 11 / Sept 4 / Sept 23 / Oct 26 rulemaking dates, not merely the docket URL for a secondary-sourced quote. The agency's page implies October 26, 2026 is also the formal rulemaking hearing date ('if the formal rulemaking hearing continues beyond October 26, 2026'), but this agent's direct fetch of coag.gov/ai/ this round did not itself state a hearing location; the previously-carried description of the hearing as being 'in Denver and by videoconference' is not repeated here because it could not be re-confirmed against the agency's own page this round. As of this check (2026-09-07), no hearing outcome or final rule has yet been posted. This remains a rulemaking-process update, not a change to the underlying statute's own May 29, 2026 enactment or January 1, 2027 operative date.