CT partnering with other states to regulate artificial intelligence

A proposed bill by Sen. James Maroney would let consumers appeal decisions made by AI.

John Craven

Apr 19, 2024, 8:42 PM

Updated 177 days ago

Share:

Ready or not, artificial intelligence is already here – and it’s becoming a bigger factor in our everyday lives.
To protect consumers, Connecticut is now teaming up with other states to get ahead of the technology. But new regulations face an uphill battle this year, as the clock quickly winds down on the 2024 legislative session.
“DOESN’T STOP AT BORDERS”
If you drive a car, own a smart phone or go to the doctor, you’re already using AI.
But a growing number of people are worried. A recent Pew Research Center survey found that 52% of Americans are more concerned about AI’s risks, than are excited about its convenience. In particular, civil liberties groups say the technology could discriminate against people based on race, religion, gender or sexual orientation.
The issue goes far beyond Connecticut.
“This is an issue that doesn’t stop at borders,” said state Sen. James Maroney (D-Milford), who has become a national leader on AI legislation.
Maroney is coordinating with states as far away as Alaska to craft new regulations.
“It is time for us to step up, and states are beginning to do that,” said Shelley Hughes, a Republican state senator from Alaska. “It would be nice if there was something that wasn’t patchwork, but that is not the case.”
The bipartisan group of lawmakers said they can’t wait for Washington to act.
“The other thing that we all know is that the federal government is not going to come up with their plan,” said Texas GOP state Sen. Giovanni Capriglione.
SWEEPING REGULATIONS PROPOSED
Here in Connecticut, Maroney is proposing sweeping new regulations. His bill would let consumers know if artificial intelligence is making major decisions on health care, employment, education, housing, utilities and criminal cases. Consumers could appeal decisions made by AI.
“We see so much potential, but there are some concerns. And as legislators, it is our job to mitigate the potential downside,” Maroney said. “The last major federal legislation dealing with the internet was passed in 1998, which is before most of the people I work with were born.”
The proposal also lifts the veil on how AI systems work – and the data they use to make decisions. Users and developers of “high risk intelligence systems” would have to disclose the specific purpose for using AI, any “known or reasonably foreseeable risks of algorithmic discrimination,” how the system was “trained” and where its data comes from.
Distributing “deep fake” adult images or campaign materials would also become a crime under the bill.
TECH INDUSTRY CONCERNS
The tech industry warns that Maroney’s bill could over-regulate a rapidly growing industry.
AI has tremendous potential for improving education, enabling creative expression, and creating new business opportunities,” Adam Hizkias with Chamber of Progress, which represents tech giants like Meta and Google, told lawmakers. “We agree that discrimination is wrong, but focusing exclusively on AI systems ignores offline discrimination. A better approach is to strengthen existing civil rights law protections in the Connecticut code to ensure that the most vulnerable members of society are protected online and offline.”
That’s one approach tech-heavy Virginia is taking. While a study group looks at possible new regulations, the state is already tweaking existing laws.
“That’s why it’s going to be really important for legislators to really understand the technical pieces of this,” said Virginia Democratic Del. Michelle Maldonaldo. “The programming, the algorithms, the large language models.”
“WORK IN PROGRESS”
As for Maroney’s bill, Connecticut House Speaker Matt Ritter told reporters that it’s “a work in progress” on Thursday.
“There is this conflict right now with DECD [Department of Economic and Community Development] on some provisions of it, and how onerous it might be on some smaller start-up companies,” Ritter said. “I’m sympathetic to that argument.”
The legislation has already undergone significant changes, including removing language allowing consumers to bring civil complaints to the state Commission on Human Rights and Opportunities. The bill is likely to see more even more changes on Monday, when the Legislature’s Judiciary Committee considers it before sending it back to the state Senate floor.
Time is running out though. Connecticut’s 2024 General Assembly session ends in less than three weeks, on May 8.
OTHER STATES: ACT NOW
Colorado, which passed one the nation’s first data privacy laws in 2021, has dealt with similar concerns. But Democratic state Sen. Robert Rodriguez said it’s important to get basic rules of the road in place now, before technology outpaces the ability to keep up with it.
“What we have proposed is similar to building a car,” he said. “The requirements are the chassis on the car, where further industry requirements can be added in the future.”
Republican-controlled Texas has created the Artificial Intelligence Advisory Council. Capriglione said tech companies should embrace common-sense regulations, to avoid lawsuits and public mistrust.
“I am incredibly dismayed to see that many of the same companies that have come into my office last year begging – asking – for AI regulation are now pushing back and fighting against that,” he said.