Deepfakes, policing, mental health and more: Lawmakers look to further regulate AI
georgeclerk via Getty Images
Utah lawmakers passed a number of artificial intelligence laws in 2025 that regulate chatbots, deepfakes, mental health services and more.
This story was originally published by Utah News Dispatch.
From reining in deepfakes to regulating chatbots used for therapy or legal advice, Utah lawmakers passed a handful of laws this year that deal with artificial intelligence.
In 2024, the Legislature took its first step in regulating artificial intelligence, passing a law that created a new state office to both regulate the emerging technology, while fostering innovation. The Utah Office of Artificial Intelligence Policy is intended to be a regulatory sandbox, inviting the private sector to work with the state’s Department of Consumer Protection to identify ways in which AI can be used for good, and bad.
In the year since the office was created, the intersection of AI and mental health has emerged as an area that shows promise, but also requires extra regulation, according to lawmakers.
Amid a shortage of mental health professionals, having an AI-powered “therapist in your pocket” is exciting, said Rep. Jefferson Moss, the Utah House majority leader — but it also raises concerns over the collection of sensitive data, advertising and deceptive practices.
That prompted Moss, R-Saratoga Springs, to sponsor HB452, which passed unanimously this session and was signed by Gov. Spencer Cox on Tuesday.
During the session, Moss highlighted two areas where AI is being used in mental health care. The first is clinician-focused, where therapists, social workers, psychologists, psychiatrists and other professionals use AI to summarize and analyze meetings with clients, create treatment plans and update health records.
The other is consumer-focused, where people can use AI for mental health interventions or practicing skills learned in therapy — like Moss said, a “therapist in your pocket.”
HB452 targets the latter, creating consumer protections around the sharing of data. Now, if an AI therapy chatbot collects personal information, it can’t sell that data or use it for anything other than the mental health services it offers.
The law also prohibits businesses that run a mental health chatbot from advertising products or services to clients in discussions with the AI, unless the chatbot “clearly and conspicuously identifies the advertisement as an advertisement,” according to the bill text.
And, the business cannot use information disclosed by a client to customize advertisements within the chatbot service.
AI chatbots will now be required to remind clients that they aren’t actually talking with a real person — and if it fails to disclose itself as AI, or violates other provisions of the law, the Utah Division of Consumer Protection can fine the operator $2,500 for each offense. The operator could also be liable for other penalties, depending on how severe the violation is.
While HB452 is arguably the largest and most impactful AI-related bill to pass this session, lawmakers passed a handful of other laws targeting the new technology.
SB180, Law Enforcement Usage of Artificial Intelligence
Sponsored by Sen. Stephanie Pitcher, D-Millcreek, SB180 is in response to the growing use of AI in writing police reports. Pitcher, a former prosecutor, said she knows of several agencies in the state that use Axon’s Draft One, which can create a report based on video or audio from an officer’s body-worn camera.
These technologies would still be permitted under the law, but law enforcement will be required to disclose the use of AI in the report. The author of the police report would also be required to sign a certificate stating that they reviewed the AI-generated report for accuracy.
The law directs all Utah law enforcement agencies to create an AI policy that outlines which technologies the agency uses, and when and where the technology is permitted.
SB180 passed unanimously and was signed by the governor on Tuesday. It will take effect on May 7.
Sponsored by Senate Majority Leader Kirk Cullimore, R-Sandy, SB226 narrows AI legislation passed last year that, in hindsight, he said was too broad.
Now, instead of requiring up-front disclosure for all AI-powered chatbots, SB226 narrows the requirement for “high risk” interactions — that includes chatbots that offer financial and legal advice, or medical or mental health care.
A chatbot is still required to disclose whether it’s AI-powered if the user asks, according to the bill. But for those “high risk” services, an AI-powered chatbot will need to disclose itself when the user starts interacting with it.
“For those types of things, the consumer should know up front that they are dealing with AI,” Cullimore said during the session.
Companies that violate the disclosure rules could be subjected to a $2,500 fine for each violation. The bill, which is awaiting the governor’s signature, passed unanimously in both the House and Senate. If signed, it will take effect on May 7.
SB271, Unauthorized Artificial Intelligence Impersonation Amendments
Sponsored by Sen. Mike McKell, R-Spanish Fork, SB271 is modeled after a Tennessee law called the “Elvis Act” that protects artists and musicians from having their work copied and commercialized.
Under current Utah code, advertisements that use a person’s identity to falsely convey an endorsement are illegal. But the law doesn’t yet expand to deepfakes, which are images, audio or videos edited or generated by AI that depict a person, real or fake.
Deepfakes are becoming more and more realistic, sometimes used to create pornographic material, re-create music or movie scenes, fabricate political endorsements, and make up videos or scenarios involving celebrities.
McKell’s law would target anyone who uses a person’s name or image without their consent to endorse a product, fundraise or solicit donations through AI, digital manipulation, animation and “any other technological means.”
The law would allow victims to file a lawsuit to reclaim money made using their image or identity. The law is awaiting the governor’s signature — if signed, it will take effect on May 7.
SB332, Artificial Intelligence Revisions
Also sponsored by Cullimore, SB332 is a simple bill that extends the lifespan of the Artificial Intelligence Policy Act. Now, the state’s Office of Artificial Intelligence Policy will remain in place until July 2027. It was signed by the governor on Tuesday.
Utah News Dispatch is part of States Newsroom, a nonprofit news network supported by grants and a coalition of donors as a 501c(3) public charity. Utah News Dispatch maintains editorial independence. Contact Editor McKenzie Romero for questions: info@utahnewsdispatch.com.