Utah singled out mental health chatbots for disclosure duties

The design

Utah's legislature passed Senate Bill 149, the “Artificial Intelligence Policy Act,” which the governor signed on 13 March 2024 and which took effect on 1 May 2024. Its enrolled text created the Office of Artificial Intelligence Policy and a general disclosure duty for generative AI used in state-regulated occupations. A year later the legislature passed House Bill 452, “Artificial Intelligence Amendments,” signed 25 March 2025 and effective 7 May 2025. Its enrolled text adds a new chapter, Title 13, Chapter 72a, “Artificial Intelligence Applications Relating to Mental Health,” aimed specifically at what it calls a “mental health chatbot.”

What the evidence says

The 2025 act's own definition is narrow and worth quoting: a mental health chatbot is AI technology that “uses generative artificial intelligence to engage in interactive conversations with a user... similar to the confidential communications that an individual would have with a licensed mental health therapist,” where the supplier represents, or a reasonable person would believe, it “can or will provide mental health therapy or help a user manage or treat mental health conditions.” The definition excludes scripted meditation apps and tools that only route a user to a human therapist. Where the 2024 act's disclosure duty covers generative AI broadly, the 2025 chapter targets one product category, closer in effect to California's and New York's companion-chatbot laws than to Utah's own general 2024 statute.

What it asks of people

Section 13-72a-203 requires a supplier to cause the chatbot to “clearly and conspicuously disclose” that it is artificial intelligence and not a human, repeating that disclosure before first use, at the start of any session after seven days of inactivity, and whenever a user asks whether they are talking to AI. Section 13-72a-202 separately bars a supplier from selling or sharing a user's input with third parties or using that input to target advertising at the user, with an exception allowing the chatbot to recommend its own service.

Privacy and safeguards

The 2025 chapter's privacy provisions are narrower than a general data-protection statute: they restrict third-party sale and advertising use of chatbot conversation data specifically, not broader collection or retention. A companion Utah Code provision, Section 58-60-118, gives a mental health chatbot supplier an affirmative defense against certain licensing claims if licensed therapists were involved in the tool's development and it was tested before release, a liability shield for suppliers rather than a guarantee to users. The statute does not itself mandate crisis-hotline referrals; that expectation appears in company policy, not in this text.

• Does a given companion or wellness app meet the statute's definition of a mental health chatbot, or fall outside on the scripted-content exception?
• Has a supplier documented the therapist involvement and pre-release testing Section 58-60-118's affirmative defense requires?
• What has the Division of Consumer Protection published about enforcement since the chapter took effect on 7 May 2025?

Utah's two acts sit in a deliberate sequence: a broad 2024 disclosure statute followed by a 2025 chapter aimed at a single, narrowly defined category of chatbot, and the definition itself determines which products the newer rule reaches.