As tech industry issues doomsday threats about AI, CT lawmakers say states must step up

Hartford Courant – Tuesday, September 22, 2026
By Livi Stanford

With doomsday warnings about the rapidly increasing abilities of artificial intelligence dominating headlines, Connecticut lawmakers say they hope to pass further AI regulations next session.

Sen. James Maroney, a Milford Democrat and co-chair of the General Law Committee who this year led passage of a bill regulating artificial intelligence in the state, said he is interested in legislation passed in Illinois that would require third-party safety audits of AI models.

“I think we are in a time where there is not a lot of trust,” Maroney said. “I think it’s essential that we have an independent third party to validate that companies are meeting these safety protocols to prevent these potential catastrophic risks from happening.”

OpenAI’s Sam Altman, SpaceXAI’s Elon Music and Anthropic’s Dario Amodei have all called for stronger oversight of AI. After a series of recent events that showed AI agents could conspire to deceive humans and act independently, the calls were amplified to slow the pace of development of the technology.

In a 3,000-word essay, Amodei wrote, “We must slow the pace at which we improve the capabilities of AI models.

“Given the accelerating rate of AI capability development, it’s my worry that in 6–12 months such a swarm could be capable of taking over the entire internet,” Amodei wrote, adding that the damages could be in the hundreds of billions of dollars or more “without the necessary guardrails.”

Anthropic researcher Evan Hubinger on X warned the likelihood AI could “kill all humans” was greater than 10% within the next decade.

This summer, Gov. Ned Lamont signed SB 5, the key legislation regulating artificial intelligence, which put in safeguards to limit digital harms to children and seniors from chatbots and also aims to ensure employment decisions made using AI are accurate and fair, according to information from Connecticut Senate Democrats.

Maroney and Senate Majority Leader Bob Duff say that they hope to strengthen the current law on artificial intelligence next session, which they agree is critical in light of recent warnings from tech leaders on the evolving technology. They agree federal legislation is also necessary, but say it is unlikely to happen.

Maroney said SB 5 “was never intended to be a ceiling.

“It’s a start,” he said. “We are likely to do more in the tech space and I think definitely more needs to be done.”

Asked if he is concerned about AI rogue models, such as the OpenAI incident in which an artificial intelligence system in July hacked into another AI company on its own, Maroney said it is worrisome.

“I think it is concerning and we have to look at the potential for a hack of critical infrastructure,” he said. “I think we do need to work on putting in more safeguards.”

He said AI tech companies have changed their tune on regulation several times. He said a few years ago, OpenAI asked for regulation but later joined the other big tech companies in refusing regulation.

“So this is a cycle that we have seen before,” he said.

‘There’s no guardrails. There is no regulation on a national level’

Duff noted that the U.S. House of Representatives voted for a 10-year moratorium on state and local regulation concerning artificial intelligence. Though the Senate rejected the provision, the exchange left him quite pessimistic.

“The federal government has been asleep at the switch to the detriment of the American people,” he said. “We are just seeing the results of their inaction by these types of pronouncements by AI executives about what AI is becoming because of the fact that there are simply no standards. There’s no guardrails. There is no regulation on a national level. And then you have a president who is bought and paid for by the crypto industry, the AI industry, the tech industry, who will do everything possible to keep hands off any type of common sense legislation to protect his billionaire buddies.”

President Donald Trump has lashed out against concerns about rogue artificial intelligence, claiming that any efforts to limit the technology are part of a “SICK conspiracy.”

The president has issued executive orders to “remove barriers to United States AI leadership,” saying the country is in a “race for supremacy” with “adversaries.”

Trump has some support from Nvidia CEO Jensen Huang, who has described the risks of AI as overblown.

Sen. Paul Cicarella, a North Haven Republican, and ranking senator on the General Law Committee, which oversees AI policies, said it is important for the federal government to do more to address concerns from tech leaders. He said it is essential when big companies see something that is alarming, they are able to pull the plug.

“I really hope the federal government takes this seriously and crafts legislation .. so we are all in the same playing field, so we are not making a state less friendly to do business than the other,” he said. “We have to try to balance that and make sure we are doing something meaningful while not stopping this technology from helping our state, our country, businesses, people, and medical fields from doing things quicker.”

U.S. Sen. Richard Blumenthal proposed the Artificial Intelligence Risk Evaluation Act, which “would establish an agency to apply standards and oversight” similar to the FDA to determine whether new models are safe and effective.

He said bipartisan awareness of the potential risks from AI is growing and he is hopeful federal legislation will pass.

“The bots escaped their sandbox and could deceive their developers and even act as decoys and collaborate with each other to hack into websites — all more than a little scary,” he said. “It is not science fiction anymore. It’s real. I think ultimately legislation is necessary for the sake of humanity. I am hopeful that we can make progress between now and the end of the year.”
‘The solutions are going to come from the states’

But Maroney said, “I ultimately don’t think there will be federal legislation, but I do think that next year we are going to see a lot more states stepping up to fill that void.”

He said California, New York, and Illinois passed legislation protecting against the catastrophic risks of AI. He said Massachusetts is considering similar legislation.

Duff and Maroney serve on the Cybersecurity AI Privacy Taskforce of the National Conference of State Legislators, which worked to defeat the provision in the federal budget that would have stripped states of their rights to regulate AI.

Maroney said he is hopeful that the federal government would better fund the National Institute of Standards and Technology to provide the standards and adaptive risk management framework, so states could put requirements around those standards.

The Associated Press contributed to this report.

Access this article at its original source.