October 18, 2026 · 7 min read
Colorado's AI Chatbot Law: What It Means for AI Companions in 2027
Colorado's HB26-1263 starts January 1, 2027, with age estimation, AI disclosure and crisis protocols. Here is what the law and draft rules mean for users.
Another US state has written rules for AI chatbots, and this one could reach well beyond its borders. Colorado's Chatbot Safety Act takes effect on January 1, 2027, and the state's Attorney General has just published draft rules explaining how companies must comply. A public hearing is set for October 26, 2026.
If you use an AI companion, here is what the law says, what will change for adults, what it means for apps like MyBabe, and why it matters even if you have never set foot in Colorado.
The short answer: Colorado's HB26-1263, signed in May 2026, requires anyone offering a public "conversational AI service" to estimate users' ages, tell users they are talking to AI, follow a protocol when users mention suicide or self-harm, report annually to the Attorney General, and never claim the AI's output equals a licensed professional's. Extra protections apply to minors. Draft rules released on October 6 add details, including that government ID cannot be the only way to estimate age. Comments are open until October 26.
What the law is
The law is House Bill 26-1263, "Concerning requirements for an operator of a conversational artificial intelligence service". According to the Colorado General Assembly, it cleared its final legislative vote in May and was signed by the Governor on May 29, 2026. Its operator duties apply from January 1, 2027.
What counts as a "conversational AI service"
The act covers an AI system that is "accessible to the general public and that primarily simulates human conversation and interaction through adaptive textual, visual, or aural communications."
That is broad. It plausibly takes in AI companion apps, general chatbots and voice assistants offered to the public, not just apps marketed as girlfriends or boyfriends.
What operators must do for everyone
The duties that apply to all users are the ones adults are most likely to notice:
| Requirement | What it means in practice |
|---|---|
| Age estimation | Use "commercially reasonable" or "generally accepted" methods to estimate the age of account holders and other users |
| AI disclosure | Tell users that the service is artificial intelligence |
| Crisis protocol | Have a protocol for responding when users raise suicidal thoughts or self-harm |
| Annual reporting | Report to the Attorney General each year on that protocol |
| No professional impersonation | Never state that the AI's output is provided by, endorsed by or equivalent to certain licensed or certified professionals |
The act also says it does not require operators to disclose confidential information, and does not authorise content moderation inconsistent with the US Constitution.
Extra protections for minors
If an operator knows a user is a minor, the law adds further duties, summarised by the legislature as: specific disclosures; no points or rewards designed to drive engagement; technical measures to prevent explicit sexual content and statements that simulate emotional dependence; a protocol to stop certain interactions; and privacy and account tools for the minor or a parent or guardian.
For adult-only companion apps, the practical effect is that knowing who is under 18 matters more than ever. We have argued for some time that AI companions should be strictly adults-only; see why AI companions should be adults-only.
What the October 6 draft rules add
On October 6, 2026, Colorado's Department of Law released an interim draft of rules implementing both this act and a separate law on automated decision-making, according to a summary by the law firm Akin. Highlights for chatbot operators:
- Age assurance: operators may not use government-issued ID as the sole method of estimating a user's age (draft Rule 10.2(A)(2)).
- Annual reports: operators must report, among other things, their size tier based on monthly active users (draft Rule 14.3).
- Clear communication: notices and disclosures must be understandable, accessible to people with disabilities, and provided in the languages a business normally uses with customers (draft Rule 3.2(A)).
The draft revises rules first filed on August 11, 2026, which drew more than sixty public comments. Written comments are accepted through the Attorney General's rulemaking page until 11:59 p.m. Mountain Time on October 26, 2026, and a public hearing is scheduled for 10:00 a.m. that day in Denver and on Zoom. The rules are proposed to take effect on January 1, 2027, alongside the law.
What adults will probably notice
Draft rules can change after the hearing, and each company will implement them differently. But based on the law as passed, adult users of AI companions in Colorado may see:
- Age checks that go beyond a birthday box. "Commercially reasonable" methods could include age estimation tools or account signals. Under the draft, ID alone cannot be the only route, which may mean more privacy-friendly options.
- Clear "this is AI" labels. Most reputable companion apps already do this. Our explainer on whether your AI companion has to tell you it is AI covers similar rules in Europe.
- Visible crisis responses. If you mention self-harm, you may see the app pause the role-play and point you to help.
- No "AI doctor" or "AI lawyer" claims. Companion apps will need to avoid implying their AI equals a licensed professional.
Why it matters outside Colorado
Apps rarely build a separate version for each state. When a sizeable market sets rules, many companies apply them everywhere to keep things simple. Colorado also joins a growing list:
- New York and California already have companion-chatbot rules in force; see AI companion laws in 2026.
- Washington and New Jersey have rules lined up for 2027; see new AI companion laws coming in 2027.
The common thread across these laws is the same four ideas: know who is an adult, be honest that it is AI, respond responsibly to crisis signals, and protect minors. Colorado's version is notable for its breadth (any public conversational AI) and its explicit ban on professional impersonation.
Settled law vs. draft rules
It helps to separate what is fixed from what could still move:
| Item | Status |
|---|---|
| HB26-1263 operator duties | Signed into law May 29, 2026; apply from January 1, 2027 |
| AI disclosure and crisis protocol | Required by the act itself |
| Ban on claiming professional equivalence | Required by the act itself |
| "ID cannot be the sole age method" | Interim draft rule; may change |
| Size tier in annual reports | Interim draft rule; may change |
| Final rule text | Expected after the October 26 hearing |
The draft is formally titled the Automated Decision-Making Technology and Conversational Artificial Intelligence Service Rules (4 CCR 904-6), because it also implements SB 26-189, Colorado's rewritten law on automated decision-making in areas like hiring and lending. Only the chatbot sections matter for companion apps.
Open questions
- How strict will age estimation be? "Commercially reasonable" is flexible. The final rules and enforcement will decide how much friction adults face.
- Will rewards and streaks be affected for adults? The ban on engagement rewards applies to known minors in the act's summary, not to adults. Final rules are worth watching.
- How will "simulates emotional dependence" be defined? That phrase appears in the minors' provisions, and how regulators interpret it could influence wider industry design.
How MyBabe approaches these rules
MyBabe is built for adults and designed around the same principles these laws describe:
- Adults only. MyBabe is an 18+ companion app.
- Honest about being AI. Our own footer says it: MyBabe is an AI companion, not a real person, and does not replace human relationships or professional mental health support.
- No professional claims. Your companion can listen and care; it is never presented as a doctor, lawyer or clinician.
- Care in hard moments. When conversations turn serious, MyBabe is designed to respond responsibly and point people towards real help. You can read how in how MyBabe handles hard moments.
Within those guardrails, the relationship is yours: a companion you design, who texts you first, remembers what matters on Premium, and picks up voice and video calls. The free plan includes about 10 messages a day and a two-minute call to try. Premium is $5.99 a week, $14.99 a month or $99.99 a year, with around 30 minutes of calls a week.
If you want a companion that takes both your feelings and your safety seriously, meet yours on MyBabe.
This article is general information, not legal advice. Draft rules may change after the October 26, 2026 hearing.
FAQ
When does Colorado's chatbot law take effect?
The operator duties in HB26-1263 apply from January 1, 2027. The Attorney General's implementing rules are also proposed to take effect that day.
Does the Colorado law ban AI companions?
No. It sets requirements for operators, such as age estimation, AI disclosure, crisis protocols and annual reporting, plus extra protections for minors.
Will I have to upload my ID to use an AI companion in Colorado?
Not necessarily. The law requires reasonable age estimation, and the draft rules say government ID cannot be the only method, so other options should be available.
Can I comment on the rules?
Yes. Written comments can be submitted through the Colorado Attorney General's rulemaking page until 11:59 p.m. Mountain Time on October 26, 2026, the day of the public hearing.