AI Regulation Tracker / New law

Connecticut's CART Act Bars AI From Being an Employer's Defense to a Discrimination Claim and Forces AI Disclosure on Mass Layoffs

Regulatory summary: The Connecticut Artificial Intelligence Responsibility and Transparency Act, enacted as Public Act 26-15 (Substitute Senate Bill 5, titled An Act Concerning Online Safety) and signed by Governor Ned Lamont on June 2, 2026, is now law with staggered effective dates. Two employment provisions bite first, on October 1, 2026: an.

Public Act 26-15, signed June 2, 2026, tells Connecticut employers that using an automated employment decision tool is not a defense to a bias complaint, and requires mass-layoff notices to state whether AI informed the decision. Those two duties take effect October 1, 2026. Broader AEDT notice rules and AI companion rules follow later.

Primary source

By Anthony Guerriero, Founder, The Leveraged Years · Reviewed by The Leveraged Years Editorial Desk · Published July 9, 2026 · Last updated July 9, 2026

Connecticut's CART Act Bars AI From Being an Employer's Defense to a Discrimination Claim and Forces AI Disclosure on Mass Layoffs regulation briefing

The Leveraged Years AI Regulation Tracker

Key takeaways

Connecticut moved from having no comprehensive AI statute to enacting one of the broadest in the country. For employers the two hard changes are that using an AEDT can no longer be raised as a defense to a discrimination complaint, and that WARN-triggering layoffs must be labeled as AI-related or not. The act separately builds an AEDT transparency regime (developer-to-deployer information sharing plus plain-language notices to workers), AI companion safety duties including a suicide and self-harm protocol and a 9-8-8 referral, generative AI content provenance duties for large providers, frontier developer whistleblower protections, and state-agency AI governance.
HR leaders and talent-acquisition teams using AI screening, ranking, or scoring tools; in-house and employment counsel; executives signing off on reductions in force; HR technology vendors selling into Connecticut; AI companion and chatbot operators; healthcare organizations relying on the clinical-support carve-out; and compliance officers at large generative AI providers.
Status: Signed into law June 2, 2026.
Before October 1, 2026, map which AI systems influence hiring, promotion, discipline, or discharge of Connecticut workers; assemble anti-bias testing evidence for each; and build a mass-layoff checklist step that records whether AI informed the decision.

Regulatory Briefing

Bindingness: Enacted state law (public act), in force on staggered dates. Public Act 26-15 is a signed statute, not a bill or a proposal. It amends the Connecticut General Statutes and creates new sections. It is binding once each section's effective date arrives. The earliest employment provisions are operative October 1, 2026. Enforcement is by the Attorney General as an unfair or deceptive trade practice under section 42-110b of the general statutes, and the act expressly provides no private right of action for its new AI sections.
Scope: Employers, AEDT developers and deployers doing business in Connecticut, operators of AI companions, large-user generative AI providers, frontier model developers, and state agencies. The employment provisions reach any employer covered by the state discrimination statute and any employer that files a federal WARN Act notice for Connecticut layoffs. The AEDT notice duties reach persons doing business in the state who develop or deploy technology that is a substantial factor in employment-related decisions. A targeted healthcare carve-out excludes certain clinical-support AI from the companion rules.
What Changed: Connecticut moved from having no comprehensive AI statute to enacting one of the broadest in the country. For employers the two hard changes are that using an AEDT can no longer be raised as a defense to a discrimination complaint, and that WARN-triggering layoffs must be labeled as AI-related or not. The act separately builds an AEDT transparency regime (developer-to-deployer information sharing plus plain-language notices to workers), AI companion safety duties including a suicide and self-harm protocol and a 9-8-8 referral, generative AI content provenance duties for large providers, frontier developer whistleblower protections, and state-agency AI governance.
Who This Affects: HR leaders and talent-acquisition teams using AI screening, ranking, or scoring tools; in-house and employment counsel; executives signing off on reductions in force; HR technology vendors selling into Connecticut; AI companion and chatbot operators; healthcare organizations relying on the clinical-support carve-out; and compliance officers at large generative AI providers.
Current Status: Signed into law June 2, 2026. In force on staggered dates. Not repealed or stayed at time of writing.
Effective Date: Section 13 discrimination-defense change and Section 26 mass-layoff AI disclosure take effect October 1, 2026. AEDT developer and deployer notice duties (Sections 7 to 12) take effect October 1, 2026 and apply to technologies deployed on or after October 1, 2027. AI companion rules (Sections 4 to 6) take effect January 1, 2027. The minors platform provisions (Section 39) take effect January 1, 2028.
Primary Source: https://www.cga.ct.gov/2026/act/pa/pdf/2026PA-00015-R00SB-00005-PA.pdf
Secondary Analysis: Office of Governor Ned Lamont press release (June 2, 2026); Ropes and Gray, Connecticut Enacts Sweeping AI Law Covering Employment, Healthcare, and Online Safety (June 2026); Future of Privacy Forum, SB 5 in Five (June 2026).
Workflow Impact: HR must inventory every AI tool that touches an employment decision, document anti-bias testing, add worker-facing disclosure language, and rewrite layoff-notice procedures so the Labor Department filing answers the AI question. Vendor contracts need clauses allocating the developer and deployer duties under Sections 8 to 10.
Operational Next Step: Before October 1, 2026, map which AI systems influence hiring, promotion, discipline, or discharge of Connecticut workers; assemble anti-bias testing evidence for each; and build a mass-layoff checklist step that records whether AI informed the decision.
Last Verified: July 10, 2026

Date	Jurisdiction	Rule	Affected professionals	Status or effective date
2026-07-09	United States (Connecticut)	Connecticut moved from having no comprehensive AI statute to enacting one of the broadest in the country. For employers the two hard changes are that using an AEDT can no longer be raised as a defense to a discrimination complaint, and that WARN-triggering layoffs must be labeled as AI-related or not. The act separately builds an AEDT transparency regime (developer-to-deployer information sharing plus plain-language notices to workers), AI companion safety duties including a suicide and self-harm protocol and a 9-8-8 referral, generative AI content provenance duties for large providers, frontier developer whistleblower protections, and state-agency AI governance.	HR leaders and talent-acquisition teams using AI screening, ranking, or scoring tools; in-house and employment counsel; executives signing off on reductions in force; HR technology vendors selling into Connecticut; AI companion and chatbot operators; healthcare organizations relying on the clinical-support carve-out; and compliance officers at large generative AI providers.	Signed into law June 2, 2026. In force on staggered dates. Not repealed or stayed at time of writing.

Source File

Instrument: Public Act No. 26-15, Substitute Senate Bill No. 5, An Act Concerning Online Safety (Connecticut Artificial Intelligence Responsibility and Transparency Act)
Authority: Connecticut General Assembly; signed by Governor Ned Lamont
Date: Signed June 2, 2026; staggered effective dates beginning October 1, 2026
Citation: 2026 Conn. Pub. Act No. 26-15 (Sub. S.B. 5), Sections 2, 4-15, 26, 38, 39
Primary Source: https://www.cga.ct.gov/2026/act/pa/pdf/2026PA-00015-R00SB-00005-PA.pdf
Corroboration: Office of Governor Ned Lamont, Governor Lamont Signs Legislation Establishing Youth Online Safety Protections, Regulations Over Artificial Intelligence (June 2, 2026); Ropes and Gray, Connecticut Enacts Sweeping AI Law Covering Employment, Healthcare, and Online Safety (June 2026); Future of Privacy Forum, SB 5 in Five (June 2026).

Frequently Asked Questions

Is the CART Act in force now?

It is enacted and partly in force. Governor Lamont signed it June 2, 2026. Its provisions take effect on staggered dates. The two key employment provisions, the discrimination-defense change and the mass-layoff AI disclosure, take effect October 1, 2026.

Does the act ban employers from using AI in hiring?

No. It does not prohibit AEDTs. It provides that using one is not a defense to a discrimination complaint, requires worker-facing disclosures for tools deployed on or after October 1, 2027, and lets a court or commission weigh anti-bias testing evidence.

When exactly do the AEDT notice duties apply to my tools?

Sections 8 to 12 take effect October 1, 2026, but the developer and deployer notice duties apply to automated employment-related decision technologies deployed on or after October 1, 2027.

Can an employee sue us directly under the act?

Not under the new AI sections. Enforcement runs solely through the Attorney General as an unfair or deceptive trade practice under section 42-110b. The act states it provides no private right of action for those sections. Underlying discrimination claims still proceed through existing law.

What is the single most urgent employer step before October 1, 2026?

Inventory every AI system that influences hiring, promotion, discipline, or discharge of Connecticut workers, gather current anti-bias testing evidence, and add a step to your layoff process that records whether AI informed the decision for the required Labor Department disclosure.