Connecticut Passes Broad AI Omnibus Bill

On May 1, 2026, the Connecticut General Assembly completed legislative passage of Senate Bill 5 (SB 5), a sweeping “online safety” and artificial intelligence measure that commentary has described as the Connecticut Artificial Intelligence Responsibility and Transparency Act.

The law takes effect in stages, with the most important employer requirements beginning in October 2026 and October 2027.

This update is applicable to Connecticut employers that use automated employment-related decision technology or may issue WARN notices, and key employer-facing provisions were set to take effect on October 1, 2026, and October 1, 2027.

What Employers Need to Do

• Inventory any recruiting, hiring, promotion, discipline, discharge, screening, interview, assessment, scheduling, or analytics tools that could qualify as an automated employment-related decision technology under the bill’s broad definition.

• Review notice templates and workflows for applicants and employees so the organization can disclose when individuals are interacting with covered automated tools and provide advance written notice before certain AI-assisted employment decisions.

• Prepare adverse-decision procedures that can explain, at a high level, the principal reason for an adverse outcome, how the tool contributed, what type of data was used, and the source of that data, while allowing an opportunity to review and correct inaccurate data and, where technically feasible, seek human review.

• Reassess vendor contracts, governance practices, and internal documentation for anti-bias testing, validation, mitigation efforts, and human oversight, as the bill made clear that using an automated tool would not be a defense to discrimination claims.

• Update WARN-related layoff notice processes so the organization can determine whether a covered layoff is related to artificial intelligence or another technological change before notice is submitted to the Connecticut Department of Labor.

Overview

• Senate Bill 5 was a broad Connecticut AI omnibus measure, not a workplace-only bill. It combined employer rules with consumer, chatbot, synthetic content, workforce development, and state AI policy provisions.

• For employers, the bill’s most important provisions regulated automated employment-related decision technology, broadly covering tools that process personal data and generate outputs that are a substantial factor in, or materially influence, employment-related decisions.

• Beginning October 1, 2027, employers using covered tools to interact with applicants or employees would need to disclose, in plain language, that the person was interacting with such technology unless that fact would be obvious to a reasonable person.

• When a covered tool would be used as a substantial factor in an employment-related decision, the employer would need to provide written notice before the decision, including the tool’s purpose, trade name, the categories and sources of personal data analyzed, and contact information.

• If the decision was adverse, the bill contemplated explanation, data review, correction, and appeal rights, including a high-level explanation of the principal reason or reasons and, where technically feasible, human review.

• The bill also would amend Connecticut discrimination law to make clear that discriminatory use of an automated employment-related decision process is unlawful and that AI use does not shield an employer from discrimination liability. Courts and enforcement agencies could consider evidence, or the lack of evidence, of anti-bias testing and similar proactive efforts.

• Senate Bill 5 would require employers issuing WARN-style notices to disclose whether layoffs relate to the use of artificial intelligence or other technological changes.

• Outside the employment context, the bill also created state AI offices, programs, and education initiatives while regulating AI companions, frontier-model reporting, and synthetic content.

• The bill’s effective dates were staggered, with many core provisions beginning October 1, 2026, some workforce and education provisions beginning July 1, 2026, AI companion provisions beginning January 1, 2027, and certain automated-employment disclosure obligations beginning October 1, 2027.

Why This Matters

Senate Bill 5 mattered because it combined broad AI oversight and compliance rules, consumer-protection, and innovation policy with employer-specific rules that would directly affect how organizations use automated tools in hiring and other employment decisions.

For employers, the most consequential point was that the bill did not simply require transparency; it also tied automated employment decision technology directly to discrimination risk and made clear that blaming the tool or a vendor would not eliminate liability.

The staggered implementation schedule also meant employers could not assume everything would begin at once, but they still needed to start planning early because the bill would require updates to notices, adverse-decision workflows, vendor management, and WARN-related practices.

Key Risks for Employers

• Failing to identify workplace tools that could qualify as covered automated employment-related decision technology because the bill’s definition was intentionally broad.

• Using covered tools in hiring or other personnel decisions without the required interaction disclosures or advance written notices.

• Lacking processes to explain adverse decisions, identify data sources, correct inaccurate data, or provide human review where technically feasible.

• Treating AI as a shield against discrimination claims instead of documenting anti-bias testing, validation, mitigation, and human oversight.

• Overlooking the bill’s separate AI-related WARN disclosure requirement when layoffs are connected to artificial intelligence or another technological change.

Additional Information

Senate Bill 5’s broader policy structure showed that Connecticut was not pursuing only restrictions; it was also pairing regulation with workforce development, education, state coordination, and innovation initiatives such as the AI Academy, Learning Laboratory, and safe-harbor programs.

Commentary also indicated that, with limited exceptions, many provisions were expected to be enforced primarily by the Attorney General as unfair or deceptive trade practices rather than through broad private-right-of-action litigation.

Source References

• Connecticut SB 5 – An Act Concerning Online Safety.

This communication is intended solely for the purpose of conveying information. The present post might incorporate hyperlinks directing readers to websites managed by third-party entities. The inclusion of any links within this communication is meant to serve as points of reference and could encompass opinion articles from various law firms, articles from HR associations, official websites, news releases, and documents of government agencies, and other relevant third-party sources. Vensure has no authority over these external websites and bears no responsibility for their content. Furthermore, Vensure does not endorse the materials present on these websites. The contents of this communication should not be interpreted as legal advice or as a legal standpoint concerning specific facts or scenarios. Nor should it be deemed an exhaustive compilation of facts potentially pertinent to federal, state, or local laws. It is strongly advised that employers solicit legal guidance from an employment attorney when undertaking actions in response to any legal updates provided. This is due to the possibility of future alterations occurring in federal, state, and local laws, regulations, as well as the directives and guidelines issued by governing agencies. These changes may transpire at any given time, potentially rendering certain portions of the content within this update void or inaccurate.