Connecticut Lawmaker Gains Prominence as States Grapple With AI

Connecticut lawmakers will consider a new artificial intelligence measure this term, crafted by a state legislator who’s part of a national movement seeking to regulate without stifling the emerging technology.

State Sen. James Maroney (D) is set to roll out legislation next week, created with the help of an informal interstate working group of dozens of legislators devoted to avoiding a patchwork of regulations tech companies warn would be overly burdensome for businesses.

The legislation (SB 2) to be unveiled next week would crack down on AI-made political deepfakes and revenge pornography, while also addressing potential job losses with online worker retraining, Maroney said.

“We’ll see a few states probably introduce the same thing this year and then other states will tinker with it and make it better,” Maroney said.

Legislators in other states are emulating what Maroney calls a balanced approach to trying to develop restrictions without hurting technological innovation. His efforts also underscore Connecticut’s growing national influence on AI policymaking amid inaction by Congress.

“He is a legislator who is willing to listen to all the stakeholders,” said Heather Morton, a policy analyst at the National Conference of State Legislatures. “He talks to everybody.”

Bringing Industry In

Maroney has leaned on task forces to help craft AI policies. It mirrors his approach to developing data privacy legislation enacted in Connecticut over the past two years after soliciting input from tech companies.

Connecticut was one of the first states to enact comprehensive rules on the collection, sales, and purchase of consumers’ private data.

Privacy groups, however, have criticized that law for a range of reasons including the lack of a private right to action allowing citizens to sue companies for violations and a requirement that consumers be able to opt out of data collection rather than making companies get their permission beforehand.

“It doesn’t require businesses to change their business practices at all,” said Caitriona Fitzgerald, deputy director at the Electronic Privacy Information Center. “It reinforces the status quo that as long as you tell consumers what you’re doing in a privacy policy, it’s OK.”

It’s likely the same concerns will surface as Maroney confronts AI.

Lobbying efforts by tech companies, aimed at minimizing restrictions on a technology, have expanded as AI could become a $1.3 trillion industry by 2032, according a 2023 Bloomberg Intelligence report.

And it’s unlikely that a private right to action, which tech companies have generally opposed for enforcing AI rules, will make it into a final bill, Maroney said.

However, he dismisses criticism that he’s too friendly to industry interests.

“If we’re not willing to compromise and you don’t get anything passed then you’re not able to help anyone,” he said.

Tech leaders from companies like Amazon.com Inc., Deloitte LLP, and Google in testimony to the Connecticut AI Working Group touted the transformative potential of AI across society in addition to their own self-described efforts to develop it responsibly.

Washington state Sen. Joseph Nguyen (D), a longtime Microsoft executive who introduced legislation in 2019 that partially inspired Maroney’s data privacy law, implored the Connecticut panel at an October meeting to avoid “overregulation that might hinder economic opportunities” through “fear-based responses, like outright bans” on AI systems, according to a meeting summary.

Getting tech companies involved in the process could be key to increasing AI restrictions this year, said David McGuire, executive director of the American Civil Liberties Union of Connecticut.

McGuire said Maroney, who also sits on a panel organized by the National Conference of State Legislatures devoted to AI issues, is “definitely a task force guy and with the AI he did bring in people from industry and I think that was a smart thing to do.”

‘Connecticut Leads The Way’

Connecticut lawmakers will likely consider the bill at a hearing before the end of the month, ahead of the end of the 2024 legislative session in May, Maroney said. A placeholder bill introduced weeks ago already has majority Senate support.

Maroney successfully sponsored a measure last year (SB 1103) to require state agencies to disclose their use of AI systems they use and conduct impact assessments, which require AI users to consider potential risks before deploying the technology. The law also established a state panel to consider additional AI rules.

“Connecticut leads the way on emerging issues, and with AI it is no different,” said US Sen. Richard Blumenthal (D) in an emailed statement while crediting Maroney’s 2023 legislation for influencing a “Bipartisan AI Framework” he has coauthored with US Sen. Josh Hawley (R).

This legislation follows the lead of a 255-page report released by the state task force by cracking down on deepfakes while avoiding sweeping bans on the technology.

Under the upcoming bill, people in Connecticut also couldn’t distribute AI-created intimate images without the permission of people in them, Maroney said.

Campaign materials created with the technology would need to include a disclosure about the use of the technology similar to a 2023 Michigan law, he added.

The final legislation will represent give-and-take from the tech industry as well input as other stakeholders like California Assemblymember Rebecca Bauer-Kahan (D), who has sponsored legislation targeting algorithmic bias that other states like Washington are now emulating.

“With Senator Maroney’s advocacy in Connecticut and our bipartisan efforts in D.C.,” Blumenthal said, “we will develop a set of straightforward safeguards that protects against the perils of AI without stifling innovation.”