Back to News
PolicyAI Understanding briefing

Colorado adopts rules banning ID‑only age checks for AI chatbots

Colorado’s Department of Law adopted 76 pages of rules that require chatbot operators to use more than government ID for age verification, mandate AI disclosures, and set safety reporting requirements, though the rules are not yet in force.

5 min readRead the linked source
Source-provided image accompanying Colorado adopts rules banning ID‑only age checks for AI chatbots
Source referenceSource recorded
Publisher
ppc.land
Source type
Linked source — primary-source status has not been established.
ContextUnderstand this in 60 seconds

Key terms

Artificial Intelligence (AI)
The broad field of building systems that perform tasks requiring pattern recognition, reasoning, language, or decision-making.
AI Safety
A field focused on reducing harmful behavior, failures, and misuse risks in AI systems.
Inference
The runtime phase where a trained model generates predictions or outputs.
Test yourselfAI Ethics Quiz

What happened

Colorado adopted detailed regulations governing two new state AI laws – one covering automated decision‑making in lending, hiring, housing and education, and the other covering consumer chatbots. The rules prohibit operators from relying solely on government‑issued ID to estimate a user’s age, require persistent AI disclosures to minors and adults, and mandate reporting of suicide‑prevention safeguards. They also set timelines for adverse‑action notices and define how continuous chatbot interactions are measured. The rules are pending an Attorney General opinion and formal publication before they become effective.

On November 19, 2026, Colorado’s Department of Law adopted a 76‑page rule set (4 CCR 904‑6) that implements two recently enacted statutes: the Automated Decision‑Making Technology in Consequential Decisions Act (SB 26‑189) and the Conversational Artificial Intelligence Services Act (HB 26‑1263, also called the "Chatbot Safety Act"). The rules are not yet effective; they require an Attorney General opinion and publication by the Secretary of State before becoming enforceable, with a likely effective date tied to the Secretary’s notice in the Colorado Code of Regulations.

For automated decision‑making tools (ADMT) used in lending, hiring, housing, education and similar consequential decisions, the rules impose a 30‑day plain‑language notice requirement when a user is adversely affected. They define when an AI output “materially influences” a decision, provide a presumption framework, and outline exceptions for tools that merely organize or summarize data without subjective . Developers must disclose the system’s role, data sources, and any known harmful uses, while downstream users remain responsible for compliance.

For consumer chatbots, the rules introduce several novel requirements. Operators must estimate users’ ages using methods other than sole reliance on government‑issued ID, provide users with a clear, persistent AI disclosure (especially to minors), and report safety metrics related to suicide and self‑harm by July 1, 2027. A continuous interaction is defined as ending after 60 minutes of inactivity, which drives the timing of required disclosures – at the start of each interaction and at least every three hours for minors, with a lighter schedule for adults. The rules also prohibit variable‑reward mechanisms that could exploit minors and set expectations for technically feasible safeguards against sexual content involving minors.

The adopted rules differ from earlier drafts in two notable ways: the effective‑date clause now defers to the Secretary of State’s publication rather than a fixed January 1, 2027 date, and the definition of a chatbot interaction now includes a 60‑minute inactivity threshold. These changes affect how companies calculate disclosure timing and compliance deadlines.

Source details: ppc.land

Why it matters

The Colorado rules represent one of the most comprehensive state‑level regulatory frameworks for both automated decision‑making tools and conversational AI services in the United States. By banning ID‑only age verification, the state aims to prevent minors from bypassing age‑based restrictions, addressing concerns that simple ID checks can be easily forged or misused. Mandatory AI disclosures and periodic reminders to minors seek to increase transparency and reduce the risk of deceptive or manipulative interactions, especially given the rise of companion chatbots that simulate emotional relationships. The reporting requirements for suicide and self‑harm safeguards create a data‑driven oversight mechanism that could inform future federal guidance. If upheld, these rules may set a template for other states and influence national policy debates on , consumer protection, and the balance between innovation and user safeguards.

Colorado’s approach blends consumer‑protection principles with technical specificity, offering a potential model for other jurisdictions grappling with AI‑driven harms. By explicitly banning ID‑only age verification, the state addresses a loophole that could allow minors to access age‑restricted AI services, a concern highlighted by recent incidents involving inappropriate content and predatory chatbot behavior.

The mandatory, persistent AI disclosures aim to combat the “deception problem” where users, especially younger ones, may not realize they are interacting with a non‑human entity. This could reduce the psychological impact of companion chatbots that simulate empathy or emotional support, aligning with emerging research on AI‑induced attachment risks.

Requiring operators to report suicide‑prevention safeguards creates a data source for regulators to assess the effectiveness of existing safety measures and to identify gaps. This could inform future federal legislation or industry standards on AI‑related mental‑health risks.

The rules’ detailed definitions of what constitutes a consequential decision and what qualifies as an ADMT provide clearer guidance for businesses, potentially reducing litigation risk and fostering more responsible AI deployment in high‑stakes domains.

Interactive Mechanism

Interactive Mechanism: How It Actually Works

Explore the underlying technology behind this development interactively.

Agent Lifecycle Stage:
1
User Intent & Planning: "Audit customer refund request #4092 and settle payment."
2
Tool Calling: Emits structured JSON call crm_get_transaction(id='4092').
3
Guardrail & Verification:Paused: High-value action requires human operator sign-off.
4
Final Settlement: Refund recorded, email receipt dispatched, and audit log stored.
Core takeaway: An AI agent is not just a language model—it is a closed loop of planning, tool invocation, and environment feedback. Production systems require self-healing retries and strict human approval guardrails.
Interactive Concept Check+10 Points
AI Ethics Quiz

Why can ethical evaluation not be reduced to one model score?

What to watch next

Stakeholders will watch for the Attorney General’s formal opinion, the Secretary of State’s publication date, and any legal challenges, especially in light of a federal court order limiting enforcement of the automated‑decision‑making portion. Companies operating chatbots in Colorado will need to adjust age‑verification processes, potentially adopting multi‑factor or behavioral methods. Monitoring how the state enforces the disclosure cadence and safety‑reporting requirements will indicate the practical impact on chatbot design and user experience. Additionally, other states may reference Colorado’s approach when drafting their own AI regulations.

The Attorney General’s formal opinion, expected shortly after the adoption, will clarify any ambiguities—particularly around the scope of “commercially reasonable” age‑estimation methods and the extent of required disclosures for hybrid AI‑human services.

Legal challenges may arise, especially given a federal court order that already limits Colorado’s ability to enforce the automated‑decision‑making portion of the law. How the state navigates this conflict could set precedents for state‑federal interactions in AI regulation.

Industry response will be critical. Companies operating chatbots in Colorado will need to adjust their age‑verification pipelines, possibly integrating behavioral analytics, device‑level signals, or third‑party verification services. Monitoring how quickly and effectively they implement these changes will indicate the rules’ practical impact.

Other states may look to Colorado’s rulebook when drafting their own AI statutes. Tracking legislative activity in neighboring states (e.g., California, New York) for references to Colorado’s framework will reveal whether this model gains broader adoption.

Related guides & quizzes

Found this useful?