The following article was submitted by Brody and Associates, LLC. It is posted here with permission. 

The Connecticut legislature passed broad artificial intelligence legislation May 11, 2026 that includes a new framework governing the use of AI in employment-related decisions.

The bill, known as SB 5, is awaiting Gov. Ned Lamont’s signature, which is expected shortly.

Once enacted, Connecticut will join a growing list of jurisdictions that are imposing transparency and accountability requirements on employers that use AI tools in recruiting, hiring, promotion, discipline, scheduling, and termination decisions.

The statute regulates what it calls automated employment-related decision technology.

In general, AEDT refers to technology that processes personal data and generates an output that is a substantial factor in an employment decision.

The definition is broad enough to potentially cover resume-screening software, applicant ranking systems, video-interview analytics, skills assessments, productivity tools, and certain workforce management platforms when those tools materially influence personnel decisions.

What Does the Law Require?

The purpose of the law is to reduce the risk that algorithmic systems will continue or worsen historic discrimination while also giving applicants and employees more visibility into how these systems are used.

One of the most important features of the new law is its notice requirement.

Beginning Oct. 1, 2027, employers that deploy AEDT intended to interact with applicants or employees must disclose, in plain language, that the individual is interacting with such technology unless it would be obvious to a reasonable person.

When the tool’s output will be used as a substantial factor in making an employment-related decision, the employer must also provide a written notice before the decision is made.

Notice must identify the purpose of the tool, the categories and sources of personal data being analyzed, how data will be assessed, and contact information for the employer.

If such employment-related decision is “adverse,” employers must provide a high-level statement disclosing the principal reasons for the decision, including “the degree to which, and manner in which” an AEDP output contributed to the decision, the type of data used, and the right to examine or correct such data.

The law does include protection for proprietary or trade secret information, but employers should not assume that vendor confidentiality excuses them from compliance.

If an employer withholds certain information regarding the AEDP based on a third party’s confidentiality claim, the employer must nevertheless disclose that the information is being withheld and identify the legal basis for the withholding.

As a practical matter, this means employers that rely on outside vendors for screening, testing, or candidate evaluation should begin reviewing vendor contracts now to ensure they can obtain the information needed to satisfy Connecticut’s notice obligations.

Anti-Discrimination and Related Obligations

The law also makes clear employers cannot avoid liability by blaming an algorithm.

Connecticut’s anti-discrimination framework will expressly provide the use of AI or automated systems is not a defense to a discrimination claim.

In other words, if an AEDT disproportionately screens out candidates or influences decisions in a way that has an unlawful discriminatory effect, the employer may still be responsible even if the challenged output came from a third-party platform.

This provision reinforces a principle regulators have increasingly emphasized nationwide: employers remain accountable for employment decisions, whether those decisions are made by people, software, or a combination of both.

What Employers Should Do Now

For employers, the immediate takeaway is AI governance can no longer be treated as an IT issue.

Human resources, legal, compliance, and procurement teams should collaborate to identify all tools used in recruiting or personnel management, assess whether those tools materially affect employment decisions, and determine what disclosures this new law may require.

Employers should also assess whether internal policies, vendor agreements, and recordkeeping practices are sufficient to support compliance.

Even companies that already use AI responsibly may need to formalize review procedures, conduct bias testing, and create documentation explaining how automated outputs are considered by human decision-makers.

Connecticut’s new law reflects a broader regulatory trend: employers may continue using AI, but they must do so transparently, carefully, and with meaningful human accountability.

About the authors: Robert Brody is managing partner at Brody and Associates, LLC, which he founded in 1997. Matthew Chiota is a law clerk at Brody and Associates, awaiting admission to the Connecticut and New York Bar associations. Contact them at [email protected] or 203.454.0560.