Connecticut Passes Law Requiring Disclosure of AI Used in Employment Decisions
Connecticut's General Assembly passed Senate Bill 5 on May 11, 2026. Governor Lamont is expected to sign it into law. The legislation imposes significant requirements on employers and technology developers who use artificial intelligence in hiring, promotion, discipline, and termination decisions.
The law defines covered technology broadly: any system that processes personal data and produces predictions, scores, rankings, or recommendations that are a "substantial factor" in employment decisions. It excludes generic tools like spreadsheets and systems used only for descriptive purposes. The definition targets predictive AI and the risk of algorithmic bias rather than generative AI systems.
What Employers Must Disclose
Starting October 1, 2026, employers must inform employees and applicants in plain language when they interact with automated employment decision technology. This real-time disclosure is not required if it would be obvious to a reasonable person that an AI system is involved.
Before making an employment decision using AI, employers must provide written notice that includes:
- The fact that the technology is being used
- The purpose and type of decision involved
- The trade name of the system
- The categories of personal data processed and how they are assessed
- The sources of that personal data
- Contact information for the employer
Employers can contractually shift these disclosure obligations to technology developers.
Bias Testing as Evidence
Connecticut's amendment to its anti-discrimination law clarifies that using an automated system is not a defense against discrimination claims. However, courts and agencies may consider evidence of bias testing when deciding liability. This includes the quality, recency, scope, and results of testing efforts, as well as how employers responded to findings.
The approach aligns with California's anti-discrimination rules and signals that employers should conduct pre-use and ongoing testing of AI tools.
Third-Party Verification Program
The law establishes a pilot program beginning July 1, 2027, for independent verification organizations. The Connecticut Department of Consumer Protection will approve up to five organizations to assess whether AI systems meet risk mitigation and safety standards.
These verification assessments can be used as evidence in civil cases but do not create a safe harbor or presumption of compliance. The program sunsets in 2030 and serves as a testbed for potential future AI auditing regimes.
Who Enforces the Law
The Connecticut attorney general has exclusive enforcement authority. Violations are treated as unfair or deceptive trade practices. There is no private right of action, though the attorney general may grant a temporary cure period through December 31, 2027.
Technology developers must provide employers with sufficient information to comply with the law's requirements, but the statute does not specify what types of information developers must create or disclose.
What Employers Should Do Now
Although the law does not mandate formal audits or certification, its structure points toward more formal third-party validation of AI systems. Employers deploying automated employment decision technology should build internal governance frameworks that include:
- Documentation of system purpose and data inputs
- AI use-case vetting guidelines and processes
- Regular bias testing
- Vendor diligence protocols
This groundwork will help employers satisfy current disclosure requirements and prepare for likely future expectations around independent review and risk assessment.
Learn more about AI for Legal professionals and AI for Human Resources.
Your membership also unlocks:
