Colorado enacts first US AI chatbot law for minors
Key insights
- HB26-1263 is the first US state law targeting AI chatbot safety specifically for minors, with enforcement beginning January 1, 2027.
- Operators, not underlying model developers, bear direct compliance liability under Colorado's new chatbot safety framework.
- Companion bill HB26-1195 prohibits AI chatbots from conducting psychotherapy sessions or generating treatment plans without human review.
Why this matters
Colorado sets the compliance floor that every AI chatbot operator serving US minors must now treat as the minimum bar, regardless of where they are headquartered. The operator-not-developer liability model will pressure platforms like Character.AI and Replika to deploy age-verification and content-filtering infrastructure they have previously delayed. The companion therapy bill signals that state legislatures are willing to regulate AI in clinical settings, directly threatening the business models of mental health chatbot companies like Woebot and Wysa operating in the US.
Summary
Governor Polis signed HB26-1263 on May 29, making Colorado the first US state to regulate AI chatbots for minors, effective January 1, 2027.
Operators must disclose their product is not human, ban sexual content for users under 18, prohibit false emotional-dependency tactics, require parental controls for under-13s, and surface crisis resources on self-harm signals. A companion bill bars AI from leading therapy sessions without human review.
Essentially: (Character.AI, Replika, major LLM platforms) have 18 months to rebuild minor-facing flows.
- Liability lands on operators, not model developers.
- The therapy bill separates AI tools from licensed clinical practice.
Colorado's enforcement record will determine whether this spreads to other states or gets preempted federally.
Potential risks and opportunities
Risks
- Character.AI and similar companion-chat platforms face an ambiguous liability window if enforcement interpretations extend to pre-2027 interactions raised in user litigation
- Mental health chatbot companies (Woebot, Wysa, Headspace) operating in Colorado face compliance uncertainty as HB26-1195's definition of 'direct psychotherapy' has not yet been set by regulation
- Operators who invest in Colorado-specific compliance infrastructure before January 2027 face stranded costs if a federal AI framework preempts the state law before enforcement begins
Opportunities
- Age-verification and parental-control infrastructure vendors (SuperAwesome, Veriff, Yoti) gain a direct enterprise sales case in every AI chatbot company serving US minors
- Law firms with AI regulatory practices (Covington, Wilson Sonsini) are positioned to capture compliance advisory work as operators race to interpret HB26-1263's 'false emotional dependency' standard
- AI companion platforms that achieve demonstrable Colorado compliance before 2027 can market that certification as a trust differentiator to parents and school districts ahead of competitors
What we don't know yet
- Whether Colorado's age-verification requirement for under-13 parental controls will mandate biometric or ID-based solutions, and who bears the implementation cost
- How 'false emotional dependency' will be defined in enforcement guidance, and whether current Character.AI memory and persona features trigger that prohibition
- Whether HB26-1263's operator-liability framework will survive First Amendment challenge before the January 2027 enforcement date
Originally reported by cohousedems.com
Read the original article →Original headline: Colorado Governor Polis Signs First US State Law Requiring AI Chatbots to Disclose to Minors, Prohibit Sexual Content, and Surface Suicide Resources