By Grace Boudjalis
BLOG OVERVIEW: Connecticut’s Artificial Intelligence Responsibility and Transparency Act (CART Act), signed into law on May 27, 2026, is one of the broadest state AI laws to date and imposes new compliance obligations on employers that use automated employment-related decision technology (AEDT) for hiring, promotion, discipline, and other personnel decisions. Employers must tell workers and applicants when they are interacting with AEDT and provide written pre-decision notices, with phased compliance deadlines of October 1, 2026 and October 1, 2027. Connecticut employers should inventory their AI tools, conduct proactive bias audits, and review vendor contracts to confirm developer disclosure obligations before the deadlines take effect.
On May 27, 2026, Connecticut Governor Ned Lamont signed the Connecticut Artificial Intelligence Responsibility and Transparency Act, known as the CART Act and Public Act 26-15, into law. This new legislation is one of the broadest laws passed by a state to regulate artificial intelligence (AI) to date. Several provisions carry direct compliance obligations for employers.
Automated Employment Decision Tools (Sections 7–12)
Any employer deploying “automated employment-related decision technology” (AEDT) in Connecticut faces new disclosure obligations. “AEDT” is defined broadly as any technology whose output (a score, ranking, prediction, or recommendation) is a substantial factor in a hiring, promotion, discipline, discharge, or renewal decision.
Specifically, employers (i.e., deployers of an AEDT) must:
- Notify employees and applicants when they are interacting with AEDT in plain language. An exception applies when it would be obvious to a reasonable person.
- Provide a written pre-decision notice before any AEDT-influenced employment decision is made, disclosing the tool’s purpose and trade name, the categories of personal data analyzed and how, the sources of that data, and employer contact information.
AI developers must supply employers with all information needed to meet these obligations, but only if the tool was marketed or contracted for employment use. Developers may also contractually assume the employer’s disclosure duties entirely, making vendor contract negotiations an important early compliance step. If trade secrets prevent full disclosure, the withholding party must notify the other and state the reason.
Employers face two compliance deadlines. The first deadline is October 1, 2026, when developer information-sharing obligations and the trade-secret carve-out go live. Critically, the law grants no private right of action. Only the Attorney General may bring suit and must first issue a 60-day cure notice for violations occurring before December 31, 2027. The second deadline is October 1, 2027, when the employer notice obligations begin applying to newly deployed AEDTs.
AI is not a Defense Against Discrimination (Sections 13–14)
Using AEDT does not shield an employer from a discrimination claim. The law amends Connecticut’s employment discrimination statutes, which cover race, age, sex, disability, sexual orientation, and other protected classes.
While the law does not specifically require bias auditing, it creates a clear incentive to conduct it, noting that courts and the Commission on Human Rights and Opportunities may consider evidence of anti-bias testing as a mitigating factor, including its quality, recency, scope, and results. This is a meaningful provision for employers and mirrors the approach taken by other states, such as California, where proactive bias mitigation efforts can factor into how liability is assessed. Employers who have invested in documented, rigorous testing of their AI tools before and during deployment are in a stronger position if a discrimination claim arises.
The Bigger Picture: Connecticut Is Watching AI’s Impact on Workers (Section 29)
Connecticut has commissioned a formal study, led by the Institute for Municipal and Regional Policy at the University of Connecticut, to track AI-related layoffs, assess impacts on entry-level jobs and underrepresented workers, and develop policy recommendations. A report based on this study is due to the General Assembly by January 1, 2027. Employers should expect that data collected now will shape future legislation.
What Employers Should Do Now
Employers with a presence in Connecticut should take several steps to begin preparing for the two compliance deadlines in the coming months. These steps include the following:
- Examine current tools to determine whether they qualify as AEDT under the statutory definition.
- Conduct proactive bias audits to surface risk early and establish documented records Connecticut courts may weigh in future discrimination proceedings.
- Review vendor contracts to confirm what information developers will provide and whether they will assume disclosure duties.
DCI will continue to monitor developments under Connecticut’s AI law. Stay tuned to the DCI Blog for updates and reach out with any questions about your AI compliance strategy.
