Utah’s Moderate Approach to AI Regulation
The train of state AI legislation keeps moving forward as many state legislatures wind down their annual sessions. Some key highlights:
Utah Gov. Cox signed a package of six AI-related bills into law, which are the subject of our primary analysis below.
Lawmakers sent deepfake bills to the governor's desks in Wisconsin (political deepfakes) and Idaho (sexual deepfakes). Dozens of additional AI bills (mostly deepfake related) are awaiting gubernatorial signatures to become law across the country.
A California regulatory board voted to proceed toward formal rulemaking on automated decision-making regulations.
And a key legislative committee approved Connecticut’s comprehensive AI bill this week but not before making a handful of substantive amendments.
Utah enacted a package of AI bills into law this week. While not revolutionary, Utah’s legislative package represents a middle ground that most state policymakers are taking on AI regulation. These bills do not establish a comprehensive framework for the development and deployment of AI (they’ll let California and Connecticut take the lead on that), but the Utah bills address pressing issues (deepfakes and consumer protections) and align current laws with the new realities that AI tools represent, with particular emphasis on protecting vulnerable populations and mandating transparency.
Utah Governor Cox (R) signed six AI-related bills into law on Wednesday. These new laws nicely illustrate the range of AI issue areas that state lawmakers have focused on this year across the country. These bills have a major emphasis on generative AI.
UT HB 148 (sexual deepfakes): Amends the Sexual Exploitation Act to include computer-generated videos.
UT HB 238 (sexual deepfakes): Amends the definition of "child sexual abuse material" to include artificially generated content.
UT SB 66 (fraudulent deepfakes): Amends the definition of “counterfeit intimate image” in the criminal code to include generated images.
UT SB 131 (political deepfakes): Requires political advertising containing synthetic media to include a disclosure. Allows a court or other sentencing body to consider the use of artificial intelligence as an aggravating factor in sentencing.
UT SB 231 (biometrics): Prohibits a governmental entity from obtaining biometric surveillance information without a warrant.
UT SB 149 (disclosure): Requires anyone who causes generative AI to interact with a person to disclose the person is interactive with AI and not a human. Creates the Office of Artificial Intelligence Policy. Establishes the Artificial Intelligence Learning Laboratory Program. Enables temporary mitigation of regulatory impacts during AI pilot testing.
Not surprisingly, four of the six bills target deepfake media. That has been, by far, the most popular topic for state legislatures to address this year, keeping us squarely in the narrowly focused phase of AI regulation. The bill limiting state use of biometric information is also a topic states had taken up before the newest AI technology burst on the scene but has received additional scrutiny since then.
The most unique aspect of the legislative package is the “Artificial Intelligence Policy Act” (UT SB 149). The first section of this law is focused on consumer protection and transparency measures. The bill language clarifies that the use of an AI system is not a defense for violating the state’s consumer protection laws. And the law requires certain licensed professionals (e.g., mental health providers) to proactively disclose when a consumer is interacting with AI technology while other professionals (e.g., telemarketers) must disclose AI use when asked by the consumer.
The second part of the law focuses on studying AI policy further and encouraging the testing and development of AI systems in the state. The law creates and grants regulatory authority to the Office of Artificial Intelligence Policy. That office is tasked with establishing an Artificial Intelligence Learning Laboratory Program. In addition to researching the potential impacts of AI, evaluating current AI policy, and making policy recommendations to the legislature, the program should “encourage the development” of AI in Utah. Finally, the office will oversee and approve temporary safe harbor protections (“mitigation agreements”) for developers, allowing applicants to test AI systems in the state within a limited scope while avoiding some regulatory restrictions.
The AI Policy Act, along with the five additional AI bills signed into law this week in Utah, threads the needle between promoting the promising benefits of AI development and mitigating the real risks and costs associated with the technology. Senator Kirk Cullimore (R), the lead sponsor of the AI Policy Act, told a legislative committee that “we believe government should have a light touch, and we want to encourage innovation in Utah."
Recent Developments
In the News
EU AI Law: On Wednesday, European Union lawmakers gave final approval to a comprehensive AI law, putting it on track to take effect later this year. The EU has taken a broader approach to regulating the development and deployment of AI than the US government is likely to take, although many aspects of the EU law are being debated on the state level in America.
Major Policy Action
California: The CPPA Board voted to take a step towards formal rulemaking on regulations for automated decision-making technology. The proposed update clarifies the contents of a risk assessment, amends considerations for impacts of consumer privacy, and considers safeguards. A final vote on whether to proceed with final rulemaking may not occur until this summer, and rules may not be finalized until 2025.
Connecticut: On Tuesday, the Joint Committee on General Law approved a comprehensive AI bill (CT SB 2) out of committee. The committee made several substantive amendments to the bill, including adding oversight from the Commissioner of Consumer Protection, requiring developers to maintain certain technical documentation of models, requiring disclosure for synthetic media, providing for protections against discrimination, and establishing grant programs for the use of AI to reduce health inequities. The bill sponsor, Sen. James Maroney (D), doesn’t anticipate the bill being included in the state budget, and added that AI is “something that we're going to have to go back to every year.”
Political Deepfakes: Lawmakers in Wisconsin passed legislation (WI AB 664) limiting the use of deepfake media during electoral campaigns, sending the bill to the governor for his signature to become law. Similar bills passed their legislative chamber of origin this week in Mississippi (MS SB 2577, MS HB 1689) and Colorado (CO HB 1147), sending those bills to the other chamber for consideration.
Sexual Deepfakes: Lawmakers in Idaho passed legislation (ID HB 575) limiting the creation and distribution of nonconsensual deepfake media, sending the bill to the governor for his signature to become law. Similar laws passed their legislative chamber of origin this week in Tennesee (TN HB 2163), Arizona (AZ SB 1336), and New Hampshire (NH SB 464), sending those bills to the other chamber for consideration.
Kentucky: On Tuesday, the Senate passed a bill (KY SB 317) that would establish property rights in an individual’s name, voice, or likeness to protect those rights from nonconsensual infringement by generative AI. The bill now heads to the House for consideration.
Arizona: On Tuesday, the Senate passed a bill (AZ SB 1599) that would make the use of AI during the commission of a crime an aggravating circumstance in sentencing. The bill now heads to the House for consideration.
Notable Proposals
Florida: An education bill (FL HB 1361) was substituted this week to create a program to award grants to school districts to implement artificial intelligence in support of students and teachers. The bill appropriates $2 million for the grants for students in grades 6-12.
Rhode Island: A proposed bill (RI HB 8058) would prohibit the use of algorithms and AI in setting residential rental prices. The FTC recently took action to combat price fixing through the use of algorithms in the rental market.