The Race — Building Democracy in the Age of AI. People race toward a shared finish line.
The Race
← Specific Events

Colorado

Scorecard

Colorado signs conversational AI service operator requirements for public chatbots

What happened

Governor Jared Polis signed HB26-1263, the Conversational Artificial Intelligence Service Operator Requirements bill. The law applies to operators of publicly available conversational AI services that primarily simulate human conversation through text, visual, or audio communications. It requires AI disclosures, commercially reasonable age-estimation methods, minor-specific protections, limits on engagement rewards and simulated emotional dependence, restrictions on certain sexual content for minors, self-harm response protocols, annual reporting to the Attorney General regarding those protocols, and limits on claims that chatbot output is equivalent to licensed or certified professional services. The law takes effect January 1, 2027.

Potential direction of power

Dispersion · Dispersion +1

The law is a modest democratic-accountability and user-protection signal. It does not govern the deepest concentration layers of AI power — compute, cloud, foundation models, distribution channels, procurement, or default platform control. But it does regulate the point where people directly experience AI as a humanlike conversational system. It gives users, parents, and regulators more leverage over whether an AI system may simulate human presence, professional authority, emotional dependence, or engagement incentives without clear guardrails.

Compliance may be easier for large firms than small developers, especially if age estimation, reporting, content controls, and self-harm protocols become expensive or legally complex. The law could unintentionally favor large platforms if implementation guidance becomes burdensome without providing clear safe harbors or scalable compliance pathways.

Window of contestability

Preserves

Positive, but narrow. The law makes a specific AI interface more visible and contestable by requiring disclosure that the user is interacting with AI, limiting professional-equivalence claims, creating minor-specific safeguards, and requiring self-harm response protocols and reporting to the Colorado Attorney General. The effect is not broad enough to treat as a major structural transparency reform.

Supporting sources

  1. denverpost.com · Supporting source 1 ↗
  2. leg.colorado.gov · Supporting source 2 ↗
  3. cohousedems.com · Supporting source 3 ↗
  4. gazette.com · Supporting source 4 ↗