Connecticut AI Laws Take Effect Oct. 1 With Chatbot Safeguards

Connecticut will implement two sweeping artificial intelligence laws on October 1, establishing some of the nation's most comprehensive AI regulations as state officials acknowledge the federal government's failure to act on the technology.

Attorney General William Tong announced the implementation timeline at a September 17 news conference, framing the state action as a necessary response to federal inaction. "We are late to this party," Tong said, adding that lacking federal leadership, Connecticut needed to take the lead on AI regulation.

Two bills address privacy and AI transparency

Public Act 26-24 amends Connecticut's existing Data Privacy Act with provisions that extend beyond AI. The law prohibits surveillance pricing, requires disclosure when facial recognition technology is used, and bans the sale of consumers' precise geolocation data.

Public Act 26-15, known as the Connecticut Artificial Intelligence Responsibility and Transparency Act (CART Act), focuses specifically on AI systems. The law targets companies offering AI chatbots to children, mandating protocols to detect and address conversations indicating self-harm risk. Chatbots must include parental controls and are explicitly prohibited from engaging in romantic or sexual interactions with minors.

State Senator James Maroney, vice-chair of the General Assembly's Energy and Technology Committee, cited research from Common Sense Media showing nearly three-quarters of teenagers have used chatbots. "It shouldn't have been necessary for us to do this, but we have to prevent the chatbots from leading children into sexually explicit conversations and inappropriate relationships," Maroney said.

Employment protections and whistleblower provisions

The CART Act requires employers to provide written notice when AI systems are used to make decisions affecting employment status. The law prohibits companies from using AI in ways that discriminate against employees.

Both bills establish whistleblower protections for individuals who report catastrophic AI risks. Any such reports must be made available to the attorney general for potential enforcement action, according to Maroney.

Why it matters

Connecticut's AI regulations arrive as states increasingly fill the vacuum left by federal inaction on technology governance. The October 1 effective date follows recent industry turbulence, including a high-profile whistleblower resignation from AI company Anthropic and reports of extensive monitoring systems tracking AI critics. For businesses operating in Connecticut, the laws create concrete compliance obligations around chatbot design, employment AI systems, and data handling—with the attorney general's office positioned to enforce violations.

Tong dismissed concerns that aggressive AI regulation might deter technology companies from Connecticut. "We have to focus on people, children, and protecting families," he said. "And yes, and by the way, protecting jobs, too."

The timing of the announcement followed AI-related developments the previous week, when an Anthropic whistleblower raised concerns about existential threats from AI and reports emerged about the company's monitoring of activists protesting AI technology.

These details were first reported by CTNewsJunkie.