SB 5: Preparing for Connecticut’s Artificial Intelligence Responsibility and Transparency Act - The National Law Review

Connecticut has enacted one of the most wide-ranging state AI laws to date. Companies that develop or deploy automated tools affecting Connecticut employees, consumers, or minors should begin assessing whether the new law applies to them, since obligations start phasing in on October 1, 2026.

Connecticut SB 5: An Act Concerning Online Safety

Effective date: Staggered, beginning October 1, 2026, through January 1, 2028

After a year of stalled negotiations over a predecessor bill in 2025, Gov. Ned Lamont signed Substitute Senate Bill 5 into law on May 27, 2026, as Public Act No. 26-15. Legislators have referred to it as the Connecticut Artificial Intelligence Responsibility and Transparency Act, or the “CART Act.” Unlike Colorado’s recently narrowed approach to AI regulation (which we covered in a separate blog post), Connecticut chose breadth over simplicity — addressing five distinct categories of AI activity in a single bill: (1) automated employment decision tools, (2) AI companions and chatbots, (3) frontier model developers, (4) generative AI content provenance, and (5) social media platforms used by minors.

Who Is Covered

The law’s reach depends on which provision of the CART Act is involved. Employers that use “automated employment-related decision technology” or “AEDT,” which is defined as “any technology that processes personal data and uses computation to generate any output, including, but not limited to, any prediction, recommendation, classification, ranking, score or other information, that is a substantial factor used to make or materially influence an employment-related decision,” must comply with new notice obligations once the AEDT is deployed in the state on or after October 1, 2027. A “substantial factor” is any output, including a “constraint, ranking, score, recommendation or classification, that meaningfully alters the outcome of an employment-related decision.”

Separately, operators of “AI companions” or systems with a natural language interface that provide “adaptive, human-like responses to user inputs” and are “able to sustain a relationship across multiple interactions” face safety and disclosure obligations beginning January 1, 2027. The law also reaches “frontier developer[s]” that train foundation models using more than 10²⁶ computing operations, providers of generative AI systems with more than one million monthly users, and operators of social media platforms that algorithmically recommend content to users under age 18.

Key Obligations Under the New Law

- Employers Using AEDT – Developers of AEDT must supply deployers, such as employers, with the information needed to satisfy notice obligations or contractually assume those obligations themselves. Beginning October 1, 2027, employers must (i) disclose in plain language when an applicant or employee is interacting with AEDT; (ii) provide written pre-decision notice disclosing the technology’s use, purpose, trade name, and the categories and sources of personal data analyzed; and (iii) provide contact information for the employer. The law also amends Connecticut’s employment discrimination statute to make clear that the use of an AEDT “shall not be a defense against a complaint alleging a discriminatory practice,” though courts have yet to test this provision.

- AI Companions – Operators must build in evidence-based protocols to detect user expressions of suicide risk, self-harm, or imminent violence, and must refer affected users to resources such as the National Suicide Prevention Lifeline. For users the operator knows or has reason to believe are minors, the law bars the companion from encouraging self-harm, discouraging users from seeking help, engaging in romantic or sexually explicit interactions, or deploying manipulative engagement tactics such as simulating distress when a user tries to end a conversation.

- Frontier Developers – “Large frontier developer[s],” meaning those with more than $500 million in annual gross revenue, must establish anonymous internal reporting channels for employees who raise “catastrophic risk” concerns by January 1, 2027, and must share updates with officers and directors quarterly. All frontier developers are barred from retaliating against whistleblowers, with violations subject to a civil penalty of up to $1,000 per violation.

- Generative AI and Minors Online – Providers of consumer-facing generative AI systems with more than one million monthly users must embed provenance data in AI-generated audio, image, and video content. These providers must use provenance data methods consistent with standards such as those from the Coalition for Content Provenance and Authenticity. Separately, beginning January 1, 2028, social media platforms (i) may not serve algorithmically curated feeds to minors without verifiable parental consent; (ii) must limit default feed access to one hour daily; and (iii) must display a prescribed warning stating that “[t]he Surgeon General has warned that while social media may have benefits for some young users, social media is associated with significant mental health harms and has not been proven safe for young users.”

- Employers Issuing Layoff Notices – Effective Oct. 1, 2026, employers issuing mass-layoff notices under the federal WARN Act must also disclose to the Connecticut Department of Labor whether the layoffs are related to the employer’s use of artificial intelligence or another technological change.

A Bipartisan Response to Washington’s Inaction

The governor, attorney general, and lead author of SB 5 developed the bill jointly, which was framed as a state-level response to a stalled federal debate over AI regulation. In announcing the signing, Gov. Lamont said the state “can no longer wait for Washington, D.C. to do the right thing and enact protections over these digital tools that give parents more control — it is time for the states to take action,” adding that the law “strikes a good balance between protecting children, ensuring workers are being treated fairly, and promoting innovation.” Attorney General William Tong was more pointed, calling the law “a major bipartisan step towards reclaiming parental control over our kids’ exposure to dangerously addictive and deeply destructive social media platforms, and an important first step towards harnessing and containing the possibilities and risks of artificial intelligence.” Sen. James Maroney, the bill’s principal author, described it as legislation that “safeguards children from the dangers of unregulated AI chatbots, shields residents from potential bias in the hiring process,” and sets the state “toward becoming the most AI-literate workforce in America.”

That last point is borne out by the bill. The law also establishes a “Connecticut AI Academy,” an AI regulatory sandbox program, and several workforce and small-business AI literacy initiatives, reflecting the Connecticut Legislature’s stated intent to pair new restrictions with investment in AI adoption.

Preparing for Compliance

With the earliest obligations taking effect October 1, 2026, companies operating in Connecticut should take the following steps to ensure compliance:

- Inventory AI systems that are used in hiring, promotion, discipline, or termination decisions, and separately track any AI companion, chatbot, or generative content tools offered to Connecticut consumers.

- Coordinate with vendors that embed AI systems in their services on document-sharing and disclosure responsibilities. Consider whether these contracts should allocate AEDT notice obligations to the developer.

- Build pre-decision notice templates and interaction disclosures for AEDT.

- Assess whether any consumer-facing chatbot could qualify as an AI companion, and if so, begin implementing self-harm detection protocols and minor-specific safeguards ahead of the January 1, 2027 effective date.

- Monitor any forthcoming rulemaking or enforcement guidance from the Connecticut Attorney General’s Office, which holds exclusive enforcement authority for most provisions under the Act.

As the provisions of SB 5 go into effect, businesses touching Connecticut employees, consumers, or minors should treat it as an immediate compliance priority.