Despite Narrow Focus, AI Bills Stall in State Legislatures
This year, several states introduced large comprehensive consumer protection AI bills, many of which have been substantially amended or failed to see much movement. However, states have also introduced smaller, more targeted AI legislation that addresses certain topics such as insurance, renting, and employment. But even these more targeted AI bills appear to be struggling to make it through the legislative process, with only a few making it to a governor’s desk.
More targeted AI legislation attempts to regulate and address specific concerns and uses of AI. Some lawmakers may have felt that these bills had better likelihood of success given their more narrow focus and specific concerns that they addressed. However, with a little more than half of the states adjourned for the year, and many more quickly reaching the end of the legislative sessions in the coming weeks, many of these bills seem unlikely to be enacted this year.
Insurance
As we have previously discussed, the insurance industry has been of interest to lawmakers seeking to regulate AI, particularly in the utilization review process. This year, lawmakers in 26 states have introduced 65 bills addressing AI’s use in the insurance industry. However, only one of these bills has been signed into law. That bill (MD HB 820) requires certain insurance carriers, pharmacy benefit managers, and private review agents to ensure any AI system was used to make an adverse decision. However, more legislation may soon be signed into law with bills in Nebraska (NE LB 77) and Texas (TX SB 815) currently waiting for their respective governor’s signature. Additionally, another two bills in Connecticut (CT SB 10) and Nevada (NV SB 128) have passed both chambers and are likely headed to their governor’s desks in the coming days.
However, given the widespread interest and current use cases for AI in the insurance industry, it is notable that these bills have not had more success in getting signed into law and many have seen very little movement. Ultimately, of the 65 bills addressing the insurance industry's use of AI, 54 have failed to make it out of their chamber of origin.
Renting
Renting is another area where there are currently specific instances of AI being utilized to assist in setting rental rates for tenants. Additionally, there are also concerns around AI being used to set rental rates, including an ongoing antitrust lawsuit brought by the Department of Justice and several states accusing a company of engaging in price fixing through the pricing algorithm in their software.
This year, lawmakers in 23 states have introduced 51 bills addressing AI’s use in establishing rental rates. However, none of these bills have been signed into law and only five have made it out of their chambers of origin.
Of those bills, only one reached the governor’s desk. In Colorado, a bill (CO HB 1004), would have prohibited the sale, distribution, or use of an algorithm to set rental rates. However, that bill was vetoed by Governor Polis (D). In his veto message the Governor stated that while he agreed that landlords should not use algorithmic pricing software to artificially inflate prices, he was concerned that the bill would also inadvertently outlaw legitimate software that could be beneficial to consumers and landlords alike.
Of the bills that have passed out their chamber of origin, three are in California (CA SB 384, CA SB 295, and CA SB 52) with the remaining two bills in Washington (WA SB 5469), which is still in session, and Hawaii (HI SB 157), which has adjourned for the year.
Employment
Employment is another area where lawmakers have signaled interest in regulating AI and where AI is already used by many employers to assist in the hiring process. This year, lawmakers in fourteen states considered 43 bills related to AI’s use in employment.
New York was the first state to attempt to regulate AI’s use in employment with two laws — one at the local level that addressed AI use among private entities in New York City, and another at the state level directed at state use of AI in hiring. So far this year, New York is again the only state to enact legislation (NY AB 433/SB 822) addressing AI in employment by amending their existing law to remove the requirement to conduct impact assessments while requiring state agencies to publish a list of AI systems that are being used.
No other bill addressing AI in employment has passed both chambers. However, the California Senate recently passed bills addressing AI’s use in hiring (CA SB 7) and workplace surveillance (CA SB 238) while in Washington the House has passed legislation (WA HB 1622) allowing public employees to bargain over the decision to adopt AI technology.
While AI remains a hot topic in state legislatures throughout the country, it appears that interest in regulating AI is not yet producing legislation that ultimately gets enacted. However, there is no sign that lawmakers are losing interest in regulating AI, only disagreement in what that regulation looks like.
Recent Developments
Major Policy Action
Colorado: Gov. Jared Polis (D) enacted legislation (CO SB 288) that establishes a cause of action against a person who discloses an AI generated depiction of another person and amends the definition of “sexually exploitative material” to include computer-generated images.
New York: A digital provenance bill (NY A 6540) was amended last week to narrow the scope. The original version of the bill required generative artificial intelligence providers to apply digital provenance data to synthetic content, but that was amended on Thursday to apply those provisions only to “synthetic content creations systems.”
Connecticut: A state budget bill proposed in the House (CT HB 7287) would include a provision to make it unlawful to disseminate sexual deepfake content. Although many bills have been introduced in the state to criminalize sexual deepfakes, it remains one of the few that have not passed any laws on the subject.
Nevada: Nevada Gov. Joe Lombardo (R) signed political deepfake legislation (NV AB 73) into law this week and had sexual deepfake (NV SB 263) legislation sent to his desk this week as well as bills regulating AI’s use in insurance (NV SB 128) and mental health (NV AB 406).
Texas: The Texas legislature sent several bills to Gov. Greg Abbott’s (R) for his signature this week, including TX HB 149, the pared-back version of the state’s comprehensive AI legislation.
Notable Proposals
Pennsylvania: House Democrats introduced a bill on Monday (PA HB 1533) that would create legal liability for certain harms caused by artificial intelligence systems. Deployers would be liable for harms from physical injuries caused by autonomous machinery, robotics, or vehicles, consumer protection violations from price-fixing or unlawful scraping of data, violation of copyright or privacy laws, algorithmic discrimination, spreading false or misleading information, or other tortious conduct.
