IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Fight Over AI Regulation Continues in Connecticut

A bill on the future of artificial intelligence in Connecticut is nearing its final form — a top priority of Democrats in the Senate — to focus on consumer protections and job creation with the evolving technology.

Gavel relating to AI regulation, laws or policies
(Shutterstock)
(TNS) — A bill on the future of artificial intelligence in Connecticut is nearing its final form — a top priority of majority Democrats in the Senate — to focus on consumer protections and job creation in the growing, historic, and sometimes scary technology.

Meanwhile, a Virginia-based trade group is trying to massively rewrite the bill and is highly critical of its current contents, which will likely be referred this week from the Senate for review by the law-writing Judiciary Committee before returning to the Senate again for debate. The General Assembly is closing in on its May 8 adjournment deadline for this year's short, budget-adjustment session.

State Sen. James Maroney, D-Milford, the co-chairman of the legislative General Law Committee who has been focused on the emerging technology in recent years, said on Tuesday that the drafting process, following a year of meetings with the AI Working Group, has been transparent. He stressed the need to protect the public from so-called sextortion, in which photos of people are doctored to depict them as naked; and deep fakes in which political enemies target candidates, such as what occurred during the New Hampshire primary to President Joe Biden.

In a recent three-page letter to Maroney and state Rep. Mike D'Agostino, D-Hamden, the other co-chairman of the General Law Committee, officials from the Arlington, Va.-based Consumer Technology Association said the bill was a threat to the industry with "significant new duties on developers and deployers of AI," and it "would effectively mandate strict new compliance obligations that would reach far beyond Connecticut."

The letter was written by Douglas K. Johnson, vice president of emerging technology policy for the 1,300-member association, and Michael Petricone, senior vice president for its government and regulatory affairs unit. "Further, we are also concerned that heavy-handed regulation of this still nascent technology will hamper innovation and investment in the market" given the costs to comply with the "burdensome, top-down regulations."

During a recent phone interview, Johnson said the association was concerned about "piecemeal" approaches in the states. "Good policy needs to happen at the federal level," Johnson said, adding that the U.S. Senate is taking a "good approach" on the future of AI. U.S. Sen. Richard Blumenthal, in particular, has focused on preventing deep fakes in the deeply divided Congress.

Since 2021, the association, which was founded 100 years ago as the Radio Manufacturers Association, has spent about $269,000 lobbying the issue in the State Capitol, according to the Office of State Ethics.

"Saying you want Congress to make regulations is saying you want no regulation," Maroney said, stressing that the current draft of the bill reflects the input of advocates, the industry and members of the working group, which finished its work in January. The bill was recently approved 22-0 in the General Law Committee.

The current version of the bill, mirroring a similar effort in Colorado, includes a new Connecticut online academy for training through Charter Oak State College. Maroney said the state now has a troubling technological divide with more than a quarter of the population without access to high-speed internet at a time when new jobs will be available for those with AI literacy. Under Gov. Ned Lamont's plans to expand broadband, communities have until the end of April to apply for grants from a pool of $41 million in federal pandemic funding to upgrade their infrastructure.

"It's going to require a different skill set, with people retraining and upscaling," Maroney said. He noted a recent report from the McKinsey & Company management consultant group that said AI technology could widen racial economic gaps, with the difference between white and Black households at $43 billion a year by the year 2045.

State Rep. David Rutigliano of Trumbull, a ranking member of the General Law Committee who has been involved in talks to redraft legislation, said in an interview on the House floor Wednesday that groups including the state Department of Economic and Community Development and the Connecticut Business & Industry Association are warning lawmakers to avoid legislation that might discourage companies from operating in the state because of new regulation and enforcement.

"You can't squash innovation and scare these companies from Connecticut," Rutigliano said. "What happens with data privacy? You might get sucked into something that you don't even know you're using. AI has been around for a long time. A lot of people don't even know you're using it. Auto-correct is using AI. It is a balancing act for sure."

Rutigliano said that issues addressing deep fakes is likely to pass, as lawmakers continue to meet privately to redraft legislation. "We're all in agreement that that's pretty important. Deep fake is a polite way of saying revenge porn and the weird stuff that's going on out there."

He said some advocates are suggesting that legislation include so-called triggers that would hold off on full adoption until surrounding states adopt similar laws.

© 2024 The Hour (Norwalk, Conn.). Distributed by Tribune Content Agency, LLC.