A closely watched Connecticut bill aimed at preventing discrimination when artificial intelligence tools make critical decisions about people is facing concerted opposition from tech industry lobbyists and the state’s governor, who vetoed similar legislation last year. Critics argue the proposal is attempting to regulate fear of the unknown and would saddle AI companies with expensive compliance burdens that could cripple the nascent industry. Backers, however, argue states must act now because the Trump administration has made clear it isn't interested in putting new guardrails around AI.
A closely watched Connecticut bill aimed at preventing discrimination when artificial intelligence tools make critical decisions about people drew concerted opposition at a hearing today from tech industry lobbyists and the state’s governor, who vetoed similar legislation last year.State Senator James Maroney, who co-chairs a multistate working group that’s proposing AI legislation across the country, is trying again this year with SB 2. But this time he’s made significant changes, hoping to address concerns that the proposal could stifle innovation and investment for the state’s AI industry. Similar bills have been introduced this year in Rhode Island, New York and Massachusetts, among several states, said Maroney, a Democrat.
Maroney’s latest bill would create liability for “high-risk” AI developers, not just those who use the tools. It would impose new reporting requirement and mandate that companies disclose how they manage the risks of algorithmic discrimination. It would also criminalize producing or disseminating deepfakes and other synthetic images that could cause harm.
It has the backing of advocacy groups such as the Electronic Frontier Foundation, although Consumer Reports argued it should go even further and close potential loopholes around the definitions of key terms, such as “high-risk artificial intelligence system” and “substantial factor,” so that companies cannot escape responsibility.
But at the first committee hearing on the bill, the tech industry and other business groups said SB 2 is still the wrong approach. Maroney’s SB 2 is trying to regulate fear of the unknown and contains provisions that would saddle companies with expensive compliance burdens, tech groups testified.
“We want to over avoid overburdening a nascent industry with onerous requirements or forcing the disclosure of proprietary information,” said Chris Gilrein, a lobbyist for TechNet. “We want to avoid mandating reporting of information to the public that has little information in their role as consumers, but creates burdens, requirements or trade secret concerns for companies. We want to leverage existing laws and frameworks and avoid layering over and creating confusion.”
Maroney wanted specific details about their concerns around definitions and proposed enforcement by the state attorney general, asking follow-up questions of nearly every witness who testified at the six-hour-plus committee hearing. Even though multiple bills were considered, debates over Maroney’s AI bill took up much of the hearing.
Backers in the state Senate said critics may argue that AI regulation should be left to the federal government. But the Trump administration has made clear it has no intention of regulating AI over the next four years, and states like Connecticut must act now to put guardrails around the technology, before it’s too late, they said.
“The window is closing on us,” Senate President Pro Tempore Martin Looney said.
The state’s Chief Innovation Officer, testifying via video from a trade mission to India, countered that it’s too soon to regulate AI, and the state’s existing antidiscrimination statutes are already sufficient.
“We appreciate the intent to protect consumers but being early in regulating a rapidly evolving, emerging category with significant economic potential presents its own risks,” said Daniel O’Keefe, commissioner of the Department of Economic and Community Development. “Rather than being a state that welcomes and supports innovation, we become the only state in the region that resists it.”
He urged state lawmakers to instead back a bill from Gov. Ned Lamont, a Democrat, who wants to take a much different approach toward regulating AI. Lamont has introduced his own bill, SB 1249, which would set up a new venture fund to invest in AI-driven startups, as well as venture funds for quantum computing and quantum materials.
Instead of enacting new antidiscrimination regulations, the governor’s proposal would amend the state’s existing unfair trade law “to make it clear that the use of AI is not a defense or omission to a claim of discrimination, closing any potential loopholes and giving residents confidence when interacting with AI software,” Lamont said in testimony to the state legislature.
The governor’s bill would also give state residents the right to sue for unfair practices because the Connecticut Unfair Trade Practices includes a private right of action.
But Maroney's proposal raised more concerns from the tech industry and its backers.
Much like the EU’s new AI Act, Maroney’s proposal is rooted in fear about eradicating hypothetical, future harms, which is a well-intentioned but misguided approach, said Adam Thierer, senior research fellow, R Street Institute, a right-leaning think tank based in Washington DC. The proposal would only allow competitors such as China to gain an edge in the AI race, he said.
“We’re setting ourselves up for real trouble,” Thierer said.
But Maroney pushed back on that comparison. “We are looking at actual harms,” Maroney said, pointing to examples of housing discrimination by AI tools. “There are instances of discrimination.”
Please e-mail editors@mlex.com to contact the editorial staff regarding this story, or to submit the names of lawyers and advisers.