After an hourslong debate, the Senate closed out its Tuesday session by passing its latest legislative attempt at state artificial intelligence regulations, despite strong objection from some lawmakers worried that the bill was overly broad, complicated and could hurt Connecticut’s economic competitiveness. 

Shortly after Hartford officials finished celebrating “AI Day” downtown, legislators in the Capitol voted 32-4 for an amended Senate Bill 5, a comprehensive AI proposal that regulates developers of “frontier” AI models, outlines the creation of a state AI “sandbox” to help companies test new technologies and creates new regulations around youth social media and AI chatbot use, among other provisions. 

“What we’re doing is we’re putting in important protections,” said Sen. James Maroney, D-Milford, the co-chair of the General Law Committee and S.B. 5’s main author.

“Sometimes these machines get things wrong,” he said.

Determining the best way to regulate artificial intelligence has been a contentious issue in Connecticut for years, with the Senate, House and governor’s office all divided on the best way to deal with a technology that is changing faster than regulations can keep up. Despite a last-minute compromise, last year’s effort to pass regulation cleared the Senate, but fell apart due to a veto threat from Lamont, who has long worried that too much regulation would hurt businesses and hamper innovation in the state. 

This session, after years of efforts to pass legislation, and as questions around AI use, intellectual property, and privacy rights become more pressing, legislators argued that a bill needs to be passed.

“I think that this will do more good than any negative,” said Sen. Paul Cicarella, R-North Haven.

S.B. 5 contains a number of AI-related regulations and provisions, including on employment-related decision-making, and AI-related tasks and programs within state agencies. 

The bill also includes efforts to expand awareness of and involvement in AI-based workforce development in the state, such as promoting the Connecticut AI Academy among the parents and guardians of state baby bond recipients and unemployed workers, offering AI education support for small businesses and expanding AI literacy programs for teachers and state workforce programs.

Previous versions of the bill were replaced with a strike-all amendment introduced by Maroney Tuesday afternoon. To help lawmakers understand the new bill before them, Maroney and Cicarella, the ranking senator on the General Law Committee, engaged in a lengthy back and forth, with Cicarella asking the committee chair to explain the majority of the bill over the course of the first hour of debate. 

Lawmakers had previously expressed concern around the size of the legislation, a 40-section omnibus bill that regulates a range of AI-related topics in the state. “There’s so much to the bill,” Sen. Henri Martin, R-Bristol, said shortly before the bill was called. “We may want to consider not doing so much and really keep it simple.” 

But legislators ultimately decided that the proposal, even with lawmakers’ limited time to review the latest changes, should move forward. 

A different approach

This year lawmakers took a different approach to AI policy, opting to target specific uses of artificial intelligence more so than specific users. The focus is largely on a targeted list of AI-related issues including data privacy, online safety of minors, and AI-related workforce development. 

The approach mostly sidesteps regulations that would create new requirements for the state’s businesses, an issue that halted previous efforts to pass an AI bill.  

The latest AI bill arrived on the Senate floor as a patchwork that resulted from folding multiple bills into S.B. 5. The bill now includes pieces of legislation that cleared other committees earlier in the session, including a provision from the Labor and Public Employees Committee that focused on automated employment decision-making. 

Senate Bill 5 also now includes portions of two Gov. Ned Lamont-backed bills, Senate Bill 86 and House Bill 5037, adding in provisions around a state-managed “regulatory sandbox” that would allow companies to test new technologies and products, as well as governor-supported regulations around youth social media use and interactions with AI chatbots.

The latter proposal has been controversial this session, with parents, technology policy groups and digital rights organizations arguing that H.B. 5037 inserts the state into conversations around youth online activity that should be left to parents. 

The digital organizations also argue that the measure’s age verification provisions introduce privacy concerns, noting that the bill will require age verification of all social media users to be effective. To accomplish that task, more data about users across the state would need to be collected, just as legislators seek to further strengthen the state’s data privacy laws and help consumers pull their information offline. 

But the bill’s supporters countered that the legislation is needed to help negate the harms of extensive social media and chatbot use on children, an issue that has become a priority for the attorney general’s office and the governor this year.

“We have to understand that there is a problem,” said Sen. Saud Anwar, D-South Windsor, after recounting the story of a teenage boy who died by suicide after being encouraged by an AI chatbot. “We need to have important protections in place.”

The latest amendments, however, have ultimately moved Senate Bill 5 closer to having a similar scale and breadth as 2025’s Senate Bill 2, the failed comprehensive AI bill Maroney previously introduced. During a brief interview with the Connecticut Mirror on Tuesday morning, Maroney said that with numerous AI bills on the Senate’s calendar and the end of the session quickly approaching, condensing the legislation was needed. 

“There’s not as much time to get things done, so we’re trying to combine where we can,” he said. 

But the approach proved to be divisive, with some GOP senators arguing that the new bill additions arrived so late that lawmakers had no chance to understand the bill they were voting on.

“I just got this bill,” said Sen. Rob Sampson, R-Wolcott, who criticized the legislation extensively Tuesday evening. “This is yet another strike all amendment. I think the third or fourth one that we’ve gotten in about 24 hours.”

Old concerns remain prevalent

Even with the changes this year, much of the debate on the bill echoed previous arguments that have been raised against the state adopting artificial intelligence legislation. 

Much of that centered on the argument that artificial intelligence is still a new technology, and that attempting to regulate it at this time is premature. Some legislators argued that even as other states move forward with legislation, Connecticut should not be trying to pass a state AI law, saying that it is up to the federal government to regulate. 

“I fear crafting legislation in the state of Connecticut with 4 million people, on a technology that is really unknown, and is evolving by literally the second,” said Senate Minority Leader Stephen Harding, R-Brookfield. “This should be done at the federal level.”

Opposing lawmakers also argued that even with the bill changes, businesses would still be negatively affected by the legislation. 

“We may be creating a roadblock that hampers business success, business innovation,” said Sen. Tony Hwang, R-Fairfield.  

The bill’s latest amendments seem aimed at getting one key party on board: Lamont. By incorporating the governor’s social media and regulatory sandbox proposals, the legislation is now much more closely tied with issues that the governor has marked as a priority. 

In a statement, the governor’s office acknowledged the bill’s upsides. “The Governor has said that the potential benefits posed by AI must be balanced with the safety of its users,” said spokesperson Cathryn Vaulman. “This bill provides helpful clarity and promotes user safety in specific use cases.”

With the latest vote, the bill now moves to the House, which opted to not take up last year’s AI proposal. 

Lawmakers say that after several failed attempts, passing the artificial intelligence bill remains necessary. “History will not look with favor on those who say that we should wait,” said Senate President Pro Tem Martin Looney, D-New Haven. “Things are changing so quickly.”