A developing trend among states is to legislate the use of employment-related decision technology—including those using artificial intelligence (AI)—to ensure that there’s a human in the loop somewhere who is aware of what’s happening. Connecticut, for example, is imposing obligations on employers, developers, and deployers regarding AI use and notification to those affected.

Frances M. Green, JD, LL.M., is an attorney at Epstein Becker Green, a national law firm focused on health care, life sciences, and employment and workforce management. She advises boards, executives, and in-house counsel on AI governance frameworks, automated decision-making compliance, and the intersection of agentic AI deployment with multi-jurisdictional employment law.

In this article, Fran explores Connecticut’s new law on automated employment-related decision technology—focusing on what employers need to know.

On May 27, 2026, Connecticut Governor Ned Lamont signed Substitute Senate Bill 5, “An Act Concerning Online Safety,” now Public Act No. 26-15 (the “Act”), enacting one of the most wide-ranging state artificial intelligence (AI) statutes to date. The Act establishes several distinct regulatory frameworks—governing automated employment-related decision technology, frontier model developers, AI companions, subscription-based AI products, generative AI content provenance, and minors’ use of social platforms—with obligations phasing in between October 2026 and January 2028.

For employers, the Act’s most immediate operational impact falls in three areas:

pre-decision notice obligations for AI used in employment decisions;
an amendment confirming that AI use is no defense to a discrimination claim; and
a new requirement to disclose, in connection with mass-layoff notices, whether reductions are AI-related.

Connecticut Public Act No. 26-15: What Employers Should Know

Automated Employment-Related Decision Technology (“AEDT”). The Act reaches any technology that processes personal data and uses computation to generate an output—a prediction, recommendation, classification, ranking, or score—that operates as a substantial factor in making, or that materially influences, an employment-related decision. “Substantial factor” means a factor that meaningfully alters the outcome of a covered decision affecting an individual in Connecticut.

Covered “employment-related decisions” include hiring, promotion, discipline, discharge, renewal of employment, and selection for training or apprenticeship, along with decisions concerning tenure or the terms, privileges, or conditions of employment. The statute expressly carves out several categories: common-utility software (word processing, spreadsheets, spellcheck, anti-malware, spam filtering, and the like); decisions producing only nonmaterial changes to tasks, responsibilities, hours, or assignments; and decisions concerning workplace health and safety, scheduling, and productivity monitoring.

Deployer obligations. An employer that puts AEDT into use in Connecticut (a “deployer”) will face two principal duties:

Interaction disclosure. Where AEDT interacts directly with an employee or applicant, the deployer must disclose, in plain language, that the individual is interacting with the technology—unless that fact would be obvious to a reasonable person.
Pre-decision written notice. Before making an employment-related decision using AEDT, the deployer must furnish the affected individual written notice identifying that AEDT has been deployed and the nature of the decision; the technology’s purpose and trade name; the categories of personal data it analyzes and how that data will be assessed; the sources of that data; and the deployer’s contact information.

Developer obligations. Developers must supply deployers with the information needed to satisfy these duties or may contract to assume the deployer’s obligations directly—a point worth attention in vendor agreements and indemnification provisions.

Effective dates. The structural provisions allocating developer and deployer duties become operative October 1, 2026. The deployer-facing interaction disclosure and pre-decision notice obligations take full effect October 1, 2027, applying to AEDT developed or deployed in the state on or after that date.

AI Use Is Not a Defense to Discrimination

The Act amends Connecticut’s employment discrimination law to specify that using covered AEDT to make an employment decision is not a defense to a complaint alleging a discriminatory practice. Courts and the Commission on Human Rights and Opportunities may, however, weigh evidence of anti-bias testing or comparable proactive efforts—considering the quality, recency, scope, and results of such testing—as a mitigating factor. While the Act’s notice provisions carry no private right of action, the existing private right of action under Connecticut’s anti-discrimination law remains fully available, including for outcomes produced by automated tools. The practical effect is to reward documented, defensible bias-testing practices while foreclosing reliance on the technology itself as a shield.

AI Disclosure in Mass-Layoff Notices

Effective October 1, 2026, employers serving written notice of a plant closing or mass layoff under the federal WARN Act must disclose whether the covered reductions are related to the employer’s use of AI or another technological change. This provision links a new substantive disclosure to an existing federal compliance obligation and warrants integration into RIF-planning checklists now, ahead of its effective date.

Frontier Developer Whistleblower Protections

Of narrower application, the Act protects employees of “frontier developers”—entities training foundation models above a specified compute threshold—who report catastrophic-risk concerns, and prohibits retaliation against covered employees engaged in protected reporting. “Large frontier developers” (those exceeding $500 million in annual gross revenue) must, by January 1, 2027, establish an anonymous internal reporting channel, provide reasonable updates on resulting investigations, and report to officers and directors at least quarterly. Violations carry civil penalties of up to $1,000 per violation, enforceable by the Connecticut attorney general. Most employers will fall outside this definition, but clients in foundation-model development should assess these obligations directly.

Enforcement and Cure

Failure to satisfy the AEDT notice requirements constitutes an unfair or deceptive trade practice enforceable solely by the Connecticut attorney general; there is no private right of action for those provisions. For violations occurring on or before December 31, 2027, the attorney general may issue a notice of violation affording a 60-day cure period before initiating enforcement, where a cure is determined to be possible. That discretionary cure window lapses at year-end 2027.

Recommended Steps for Employers

Given the staggered runway, employers with Connecticut operations or applicants may wish to begin now to:

Inventory AI and automated tools used to make or materially influence employment decisions, distinguishing covered AEDT from carved-out utilities and monitoring functions.
Review and build out interaction-disclosure and pre-decision notice procedures against the October 1, 2027, deadline, including the specific content elements the Act requires.
Assess whether bias auditing of covered tools is warranted to mitigate disparate-impact exposure, and document the methodology, scope, and results—evidence the statute expressly permits as mitigation.
Examine developer and vendor contracts to confirm that the information needed for deployer notices will be furnished, and to allocate compliance responsibility and risk.
Integrate the WARN AI-disclosure question into RIF and layoff-planning workflows ahead of October 1, 2026.
Confirm whether an AI governance policy is in place and, if not, consider implementing one as a documented foundation for compliance.

Notably, the Connecticut WARN AI-disclosure provision and others making their way through the states reflect a developing trend toward treating AI-driven workforce reductions as a distinct category of regulatory concern—one that may warrant proactive attention to how employers document the role of AI in restructuring decisions.

About the Author

Frances M. Green, JD, LL.M., is an attorney at Epstein Becker Green, where she advises boards, executives, and in-house counsel on AI governance frameworks, employment law compliance, and privacy strategy. She holds the AIGP, CIPM, and IAPP certifications and counsels organizations navigating the intersection of agentic AI deployment, automated employment decision tools, and multi-jurisdictional regulatory compliance. She is a regular voice on the evolving federal-state tension in AI regulation; learn more about Epstein Becker Green's artificial intelligence practice.

Epstein Becker Green Staff Attorney Ann W. Parks contributed to the preparation of this article.