Quick Hit

Connecticut has become the latest state to enact legislation regulating the use of artificial intelligence as it relates to employment.  On May 27, 2026, Connecticut Governor Ned Lamont signed Senate Bill 5 (“SB 5”), formally titled the “Connecticut Artificial Intelligence Responsibility and Transparency Act,” into law.  The new law applies to employees and job applicants in Connecticut and imposes new obligations on employers doing business in Connecticut who deploy “automated employment-related decision technology.”

Key Takeaways

The enactment of SB 5 reflects the rapidly expanding patchwork of AI regulation at the state level.  Although the law is not as broad as the recently enacted AI law in Colorado, SB 5 does require employers doing business in Connecticut, as well employers who have employees in Connecticut or who accept job applications from Connecticut residents, to comply with various notice requirements beginning October 1, 2027.  As discussed below, the new law requires impacted employers to take steps to ensure compliance.

More Detail

The new law regulates “automated employment-related decision technology” (“AEDT”).  The law defines AEDT broadly 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.”  The law excludes, among other things, word processing, spreadsheet, map navigation, web hosting, data storage, anti-virus or anti-malware and “any system or service that is used incidental to making an employment-related decision” or “any information that is purely descriptive, diagnostic or statistical in nature.”

Notice Requirements Regarding Employer Use of AEDT

Beginning October 1, 2027, employers that deploy covered AEDT that is “intended to interact” with employees and/or job applicants must disclose to such individuals that they are interacting with AEDT in “plain language.”  No such disclosure is required if a “reasonable person would deem it obvious that” they are “interacting with” AEDT. 

If an employer deploys AEDT “to generate any output for the purpose of making, or as a substantial factor in making, an employment-related decision,” the employer must adhere to additional notice requirements before such decision is made.  “Substantial factor” is defined as “a factor, including, but not limited to, a constraint, ranking, score, recommendation or classification, that meaningfully alters the outcome of an employment-related decision concerning an individual in the state.”  Because the statute does not define what it means for an AI output to “meaningfully alter” the outcome of an employment-related decision, employers should expect questions regarding the scope of the “substantial factor” standard and should monitor future guidance or enforcement developments from Connecticut regulators.

Specifically, the employer must provide the individual with written notice disclosing:

• That the employer is using AEDT;
• The purpose of the AEDT and the nature of the employment-related decision;
• The trade name of the AEDT;
• The categories of personal data that the AEDT will analyze or process and how the personal data will be assessed in reaching a decision;
• The sources of the personal data; and
• Contact information for the employer.

The law does not require the disclosure of trade secrets otherwise protected by state or federal law.  However, if an employer withholds information on the basis that it is a trade secret, then the employer must send notice to the individual from whom the information is being withheld disclosing the fact of the withholding and the legal basis for such withholding. 

Any violation of these provisions is considered an “unfair or deceptive trade practice” enforceable by the Connecticut Attorney General.  For violations that occur on or before December 31, 2027, the Connecticut Attorney General may issue a “notice of violation.”  If an employer fails to cure the violation within 60 days of receipt, the Attorney General may bring an enforcement action.  The law is explicit that it does not provide a private right of action for any violation.

Amendments to the Connecticut Fair Practices Act

SB 5 amends the Connecticut Fair Practices Act—the state’s anti-discrimination law—to expressly provide that, beginning October 1, 2026, an employer’s use of AEDT is “not a defense to a complaint alleging a discriminatory practice.”  Put differently, employers cannot shield themselves from liability for discrimination claims by stating that they relied on AI. 

Importantly, however, the Connecticut Commission on Human Rights and Opportunities or a court “may consider evidence of anti-bias testing or similar proactive efforts to avoid the discriminatory practice” including “the quality, efficacy, recency, and scope of such testing or efforts” and an employer’s response to such testing results.  This provision gives employers an incentive to conduct and document anti-bias testing (or utilize products where the developer has done so) before deploying AEDT.

WARN Act Disclosure Requirement

SB 5 also provides that, beginning October 1, 2026, employers issuing written WARN notices must disclose to the Connecticut Department of Labor whether the layoffs that are the subject of the notice are “related to the employer’s use of artificial intelligence or another technological change.”

Practice Pointers for Employers

Employers should consider taking the following steps to prepare for SB 5, ideally with the assistance of counsel:

• Know your AI tools.  In order to ensure compliance, employers should have a firm grasp on which AI tools they use, what data they use AI to process, what outputs AI is generating, how AI outputs are used by human decision-makers, and whether AI outputs “meaningfully alter” employment-related decisions. Distinguishing between AI tools which merely assist with administrative tasks and AI agents that actually “meaningfully alter” the outcome of employment-related decisions will likely become one of the most important compliance questions under SB 5. If AI tools do not “meaningfully alter” employment-related decisions as the law defines them, then SB 5’s requirements will not apply.  Employers should expect that the analysis will be fact-specific, and that an AI tool’s label will matter less than how it is actually used.  Employers should begin by identifying automated tools that relate to hiring, promotion, compensation, performance management, workforce planning, or other employment decisions, focusing not on whether a product is marketed as AI but instead on what the tools actually do. To determine how these tools function in practice, employers should investigate how the tools are being used by their employees.
• Review Developer Responsibilities.  SB 5 provides different responsibilities for “developers” of AEDT (organizations that build AEDT) and “deployers” (those who put AEDT into use – in most cases, employers).  Under the law, developers of AEDT that is in use in Connecticut on or after October 1, 2027 are required to provide employers of their AEDT “all information” that such employer requires to be able to comply with the law where the AEDT is to be used to “materially influence” an employment-related decision.  Under the law, employers may contract with such developers whereby the developers assume the employer’s notice obligations under the law, but the contract must be explicit about which of the employer’s duties the developer is assuming.
• Prepare for Implementation.  SB 5 imposes new notice and disclosure requirements.  Employers should use the lead time before the law takes effect to review existing AI practices and develop procedures for providing required notices.  In addition, because evidence of anti-bias testing and other proactive measures may be considered when evaluating discrimination claims, employers should consider conducting and documenting such testing before deploying AEDT, or consult with their deployers on such testing.