Connecticut’s new AI law creates compliance obligations for employers

Artificial intelligence is rapidly becoming embedded in the modern workplace. Employers increasingly use AI-powered tools to recruit applicants, screen resumes, evaluate candidates, assess employee performance, forecast staffing needs, and support personnel decisions.

As the use of this technology expands, lawmakers have become increasingly focused on ensuring that AI tools do not create unfair or unlawful employment outcomes.

Connecticut is the latest state to enact legislation addressing the use of AI. The Connecticut Artificial Intelligence Responsibility and Transparency Act establishes a framework governing the development and use of certain AI tools and directly affects employers who use automated tools in making employment-related decisions.

Most provisions of the law will take effect on October 1 of this year. With limited exceptions, the notification requirements (described below) will not take effect until October 1, 2027.

What the new law regulates

The Act applies to all automated employment-related decision technology. “AEDT” is any technology that processes personal data and uses computation to generate an output that is “a substantial factor used to make or materially influence an employment-related decision.” 

Under the law, a “substantial factor” is any factor that “meaningfully alters the outcome of an employment-related decision concerning an individual” in Connecticut.

An “employment-related decision” is a decision to recruit, hire, promote, discipline or discharge an employee; renew employment; select an employee for training or apprenticeship; or determine tenure, terms, privileges, or conditions of employment.

Notice and transparency requirements

As of October 1, 2027, entities that deploy automated employment-related decision technology in Connecticut will be required to inform employees and applicants when AI tools are used.

First, employers must disclose to employees and applicants that they are interacting with AEDT, unless it is obvious that they are doing so. 

Second, employers who use AEDT to make employment-related decisions, or who use the information generated by AEDT as a substantial factor in making employment-related decisions, must provide certain disclosures before the decision is made:

• The employer’s contact information.
• The trade name of the AEDT.
• The fact that the employer is using AEDT.
• The purpose of AEDT use and the nature of the employment decision (for example, whether it is a hiring decision).
• The categories of personal data that the AEDT will analyze and how it will be assessed in reaching a decision.
• The sources of that personal data.

An employer can also transfer these responsibilities to its AEDT vendor via contract.  

The Connecticut Attorney General is empowered to pursue claims for unfair or deceptive trade practices for failure to provide the required disclosures. There is no private right of action available under the Act.

In addition, employers undergoing a plant closing or mass layoff covered by the federal Worker Adjustment and Retraining Notification Act must inform the Connecticut Department of Labor if the plant closing or mass layoff is related to the use of AI or “another technological change.” These WARN-specific provisions take effect on October 1, 2026.

Employment decisions remain the employer’s responsibility

The Act also addresses the interaction between AI tools and Connecticut’s anti-discrimination laws. 

Employers should be particularly mindful that the use of AEDT may create disparate outcomes affecting protected groups. Even if the employer has no intent to discriminate, AEDT use may create legal exposure if its assessment is biased.  An employer’s use of AEDT is not a valid defense under the law.  However, while anti-bias testing is not required (unlike New York City law), courts may consider mitigating evidence of an employer’s anti-bias testing or similar proactive efforts to avoid discrimination, including the following:

• The quality, efficacy, recency and scope of such testing.
• The results of the testing.
• The employer’s response to the information obtained.

Looking ahead

Employers should not wait for a regulatory inquiry or lawsuit before they evaluate their use of AI. Organizations that proactively assess their use of AI, implement governance controls, and create oversight mechanisms will be better positioned to navigate the increasingly complicated legal landscape.

If you need assistance in complying with the Connecticut law or establishing best practices, please contact Chris Deubert, John MacDonald, Jonathan Persky, or me.