Utah’s AI disclosure duty is not a universal rule that every chatbot in the state must label every message. The wording changed in 2025, and the type of interaction matters.
Know which version you are reading
Utah enacted an Artificial Intelligence Policy Act in 2024 and adopted S.B. 226 in 2025, effective May 7, 2025. The newer text draws an important distinction between a supplier’s consumer transaction conversation and AI interactions used to provide regulated professional services. Do not copy an older 2024 summary as the current obligation without comparing enacted amendments.
Consumer transaction: disclosure when clearly asked
The amended provisions call for disclosure when a person asks clearly and unambiguously whether a generative AI interaction is with a human. Businesses should train support systems to respond plainly and avoid evasive branding that masks the nature of the system. A blanket statement that every response must proactively announce AI participation goes beyond the narrow cited consumer clause.
Regulated occupations and high-risk interactions
Services in a regulated occupation face more specific conditions where the AI interaction is high risk. The law describes prominent disclosure before written interactions and at the start of verbal exchanges, alongside continued professional requirements. A practitioner cannot bypass ordinary professional rules merely by assigning client communications to a model; classification and the precise statutory definition still matter.
Design a verifiable disclosure flow
Keep a clear “AI assistant” identifier, a human escalation path and a test case in which a consumer explicitly asks if a human is replying. Verify the response and preserve a date-stamped record of the deployed configuration. For regulated services, identify the licensed professional responsible and test the pre-interaction notice; provide an accessible equivalent for callers with disabilities.
Limitations and alternatives
The law’s reach, exemptions, enforcement and statutory definitions need direct reading of the enacted Utah text and any later amendments. These controls are a product-design starting point, not a legal safe harbor. Organizations outside Utah should follow their own applicable laws rather than assuming this rule governs all US consumer AI.
Sources & Further Reading
- S.B. 226 (2025), Artificial Intelligence Amendments — Utah Legislature.
Jurisdiction/scope: Utah, United States; 2025 amended law. This is educational information, not legal, medical or other professional advice. No hands-on product testing is claimed.