Source: JD Supra

On May 11, 2026, the Connecticut General Assembly passed Senate Bill 5 and Governor Lamont is expected to sign it into law. The statute creates a comprehensive online safety framework with specific requirements for Automated Employment‑related Decision Technology (AEDT), defined broadly as systems that process personal data and produce outputs that are a “substantial factor” in employment decisions such as hiring, promotion, discipline, or termination. The law focuses on predictive AI and algorithmic bias, expressly excluding generic software used incidentally, and notes it will likely encompass generative AI when applied in ways that could cause discriminatory harm or other covered injury. The measure adopts a disclosure‑focused approach drawing on elements of California and EU rules, encouraging—but not mandating—substantive pre‑use design or audit obligations, and creates a program for third‑party risk assessments without granting broad evidentiary defenses for audited models.

Effective October 1, 2026, the law amends Connecticut’s anti‑discrimination framework so use of an automated system is not a defense to a discrimination claim, while evidence of bias testing and the quality, recency and response to such testing “may be considered” by courts or agencies. Developers and deployers receive divided responsibilities: deployers (employers) carry the primary compliance burden while developers must supply sufficient information to enable compliance when the technology is marketed to materially influence employment decisions. The statute requires real‑time plain‑language notice when individuals interact with AEDT and a pre‑decision written notice disclosing the fact of use, purpose and decision type, trade name, categories and sources of personal data, how data are assessed, and contact information for the deployer. A trade secret safe harbor permits withholding legally protected information so long as recipients are affirmatively notified that information is withheld on that basis. A pilot for up to five independent verification organizations begins July 1, 2027, under the Connecticut Department of Consumer Protection, requires state memoranda of understanding, and sunsets in 2030; approved organizations issue verification assessments with limited legal effect and do not create presumptions of compliance or formal certifications. Enforcement is through the Connecticut Unfair Trade Practices Act (CUTPA), pursued exclusively by the Connecticut attorney general, with no private right of action and an AG‑discretionary temporary cure period through December 31, 2027. The law does not mandate formal audits or certifications but signals a regulatory trajectory toward third‑party validation; employers are advised to document system purpose, data inputs, bias testing, and vendor diligence to meet disclosure requirements and anticipated expectations. The update includes a disclaimer that it is general in nature and not a substitute for specific legal advice.