Utah's AI disclosure duty: what it means for hiring

Utah's generative AI disclosure duty now lives in Utah Code Title 13, Chapter 77, effective May 7, 2025 and amended again in the 2026 General Session. It requires a business using generative AI to tell a person they are talking to AI, not a human, if the person asks. It requires proactive, unprompted disclosure only from someone in a licensed, state-regulated occupation using generative AI for a high-risk interaction. Whether an employer's AI-driven hiring chatbot or screening interview falls inside this chapter at all is a real, unresolved question, not a settled yes.

How Utah got to Chapter 77

Utah's original Artificial Intelligence Policy Act (SB 149, 2024) took effect May 1, 2024 and built its generative AI disclosure duty into Title 13, Chapter 72. In 2025, SB 226 tightened and restructured that framework: it repealed the general disclosure section in Chapter 72 and re-enacted a revised version as its own chapter, Chapter 77, "Generative Artificial Intelligence — Consumer Disclosures and Enforcement," effective May 7, 2025. Chapter 77 was itself amended again in the 2026 General Session. The current text is on le.utah.gov. Chapter 72 still exists and covers other AI policy matters, including the state's AI learning laboratory program; it is not the chapter to cite for the disclosure duty anymore.

The disclosure rule itself, in the statute's own structure

Section 13-77-103 sets two different duties, and they are easy to conflate:

A "regulated occupation" is defined in section 13-77-101(8) as one regulated by the Department of Commerce that requires a license or state certification to practice, for example, a licensed therapist, attorney or financial professional. Recruiting and HR are not licensed occupations in Utah, so the proactive duty in subsection (2) is very unlikely to reach an ordinary hiring process on its own terms. The on-request duty in subsection (1) is the one that could matter for hiring, and it turns entirely on whether a job application is a "consumer transaction."

The open question: is applying for a job a "consumer transaction"?

Chapter 77 borrows its definition of "consumer transaction" from the Utah Consumer Sales Practices Act, section 13-11-3: broadly, a sale, lease or other transfer of goods, services or property to a person for personal, family or household purposes, or certain business-opportunity arrangements. A job application is not obviously a purchase of goods or services by the applicant; it is closer to an offer of the applicant's own labor to the employer. We found no Utah statute, rule or published Division of Consumer Protection guidance stating whether an applicant interacting with an employer's AI screening tool, chatbot or interview system is a "consumer" in a "consumer transaction" under this chapter. Do not assume either answer. Two practical postures follow:

• The cautious reading. Treat any AI tool that interacts directly with a candidate, such as a chat-based screening assistant, as if the on-request disclosure duty applies, and answer honestly if asked whether AI is involved. This costs little and avoids the argument entirely.
• The narrow reading. Treat hiring as outside "consumer transaction" until a court, the Division, or the Legislature says otherwise, and rely on the fact that Utah, unlike Illinois or NYC, has not enacted a hiring-specific AI transparency law at all.

Either way, a tool that only transcribes what was said in an interview, without itself generating conversational responses to the candidate, is not "generative artificial intelligence" under section 13-77-101(4) in the first place, since that definition requires the system to simulate human conversation and generate non-scripted outputs. The distinction that matters is whether the tool talks back to the candidate, not whether it processes audio.

The safe harbor

Section 13-77-104 gives suppliers a safe harbor from enforcement under section 13-77-103: if the generative AI clearly and conspicuously discloses, at the outset of the interaction and throughout it, that it is generative AI, is not human, or is an AI assistant, the person is not subject to an enforcement action for a violation of the disclosure section. For an employer already inclined toward the cautious reading above, building that standing disclosure into any AI chat interface a candidate interacts with is the simplest way to close the question rather than debate it.

Liability, enforcement and penalties

Section 13-77-102 removes a specific defense: it is not a defense to a consumer-protection violation that generative AI made the statement, took the action, or was used in furtherance of the violation. Practically, an employer cannot point to the AI vendor's model as the responsible party for a misleading statement to a candidate; the deploying business remains answerable the same as if a person had said it.

Section 13-77-105 gives enforcement to the Division of Consumer Protection, with the Attorney General as its counsel. The division director may impose an administrative fine of up to $2,500 per violation; a court in a division-brought action can declare a violation, issue an injunction, order disgorgement, and impose a further fine of up to $2,500 per violation, plus attorney fees and costs to the division. Violating a resulting administrative or court order carries its own civil penalty of up to $5,000 per violation. There is no private right of action under Chapter 77 itself; section 13-77-106 preserves whatever other remedies exist under state or federal law.

What Utah has not done, compared with other states

Utah has not enacted anything resembling Illinois's video interview or Human Rights Act notice duties, Colorado's deployer obligations for consequential automated decisions, or NYC's independent bias audit and posting requirement. See AI hiring laws by state for the wider map, Colorado's AI law for employers for a state that does regulate hiring-specific automated decision tools directly, and California's AI hiring regulations and Texas's AI hiring law for two very different approaches to the same problem. Utah's Chapter 77 is a consumer-disclosure statute that happens to be the closest thing the state has to an AI transparency rule touching hiring, not a hiring statute in its own right.

A practical checklist

• List every point in your hiring process where a candidate interacts directly with an AI system that talks back to them, such as a chat-based screener, rather than a tool that only records or transcribes.
• For each one, add a standing disclosure that it is AI, shown at the start of the interaction and reachable throughout it, to fall inside the section 13-77-104 safe harbor regardless of how the "consumer transaction" question is eventually resolved.
• Train recruiters and hiring managers to answer honestly and immediately if a Utah candidate asks whether they are talking to AI, anywhere in the process.
• Do not rely on a vendor's own disclosure language without checking it says the tool is AI, is not human, or is an AI assistant, since that is the specific wording the safe harbor references.
• Keep this separate from any biometric, recording-consent or discrimination analysis for the same tool; Chapter 77 covers disclosure only.
• Revisit this page's reading of "consumer transaction" if the Division of Consumer Protection issues rules or guidance, since section 13-77-104(2) gives it rulemaking authority over what counts as adequate disclosure.

Common mistakes