Several States Preparing Comprehensive AI Bills for 2025
Connecticut, Texas, New Jersey and California are among the states preparing to advance comprehensive AI legislation in 2025, according to lawmakers and stakeholders.
Connecticut Sen. James Maroney (D) told us in a recent interview he plans to introduce SB-2, his comprehensive AI bill, next year. Maroney wrote the comprehensive privacy bill that became Connecticut law in 2022 after three years of effort (see 2205110049), and 2025 will mark the third year he has attempted to gain approval for a comprehensive AI bill.
SB-2 targets discriminatory, algorithm-based decisions that impact people's lives. The Senate approved the bill in 2024, but the House didn’t put it up due to a veto threat from Gov. Ned Lamont (D).
Maroney chairs the Multistate AI Policymaker Working Group, a coalition of more than 200 state legislators from 47 states. In collaboration with the Future of Privacy Forum (FPF), Maroney recently convened the organization. Its goal is establishing a basic legislative framework that can be replicated across state lines, allowing for more seamless compliance, he said.
“Most of us are part of this group because we don’t have faith in the federal government to act quickly enough,” he said. "Alone you go fast. Together you go far. And we know we have a long way to go to make sure that we’re providing protections from AI that will ultimately allow us to yield the full benefits of AI.”
Maroney has worked closely with Colorado Senate Majority Leader Robert Rodriguez (D), who led comprehensive AI efforts that resulted in a 2024 law. SB-2 targets automated decisions related to criminal justice, education, employment, financial service, government service, health care, housing, insurance and legal service. Maroney said the bill includes some Colorado provisions and data transparency provisions from California’s AB-2013. He spoke positively of provisions in Minnesota’s Consumer Data Privacy Act that allow consumers to opt out of AI-decision-related data processing and Illinois’ AI discrimination bill, HB-3733.
Maroney said SB-2 is “nothing like” California’s SB-1047, an AI liability bill that Governor Gavin Newsom (D) vetoed in September (see 2409300011). Maroney said he’s watching what happens in California, but “I want to focus on the known harms.”
In New Jersey, stakeholders are watching Senate Majority Whip Troy Singleton's (D) legislative efforts. He introduced SB-1438, which targets the automated decisions of government agencies, and SB-3742, which requires that companies report impact assessments to the government.
FPF Deputy Director-U.S. Legislation Tatiana Rice told us she will be monitoring New Jersey, Connecticut, Texas and California, among other states, in 2025. Singleton’s SB-3742 drew some inspiration from California’s SB-1047, she said. The Maroney-led working group facilitates legislators trading information and hopefully reaching a “common understanding” about comprehensive legislation, she said.
Said Maroney, “Our goal is to incorporate as many voices as possible so that we can develop regulation that will be pro-innovation.” Indiana, Alabama and North Dakota are the only states that haven’t participated in the working group.
Texas Rep. Giovanni Capriglione (R) this week released a draft proposal for a comprehensive AI bill. His legislation targets algorithmic discrimination and, like Connecticut's, it highlights decisions related to criminal justice, education, employment, financial services, government services, health care, housing and insurance.
NetChoice said in a statement that Capriglione’s draft proposal is a “good start,” but “contains a few concerning provisions that would stifle AI development and deployment in Texas.” Companies using AI tools already “comply with numerous existing laws against discrimination, bias, and abuse, across applications like healthcare, finance, and education.”
BSA | The Software Alliance said in a statement that state lawmakers introduced nearly 700 AI-related bills in 2024, and momentum is expected to continue into 2025. BSA will “urge lawmakers to focus on high-risk uses of AI and to drive toward a consistent and harmonized approach to legislation, rather than pursue divergent or conflicting models.”