Connecticut Requires New Disclosures for AI Used in Hiring and Employment Decisions
Connecticut has passed legislation regulating how employers use artificial intelligence in recruiting, hiring, promotions, and terminations. Governor Ned Lamont is expected to sign the Connecticut Artificial Intelligence Responsibility and Transparency Act (SB 5) into law, making the state the latest to impose transparency and anti-bias requirements on automated employment systems.
The law takes effect October 1, 2026, with staggered implementation dates through 2027. HR leaders need to start planning compliance now.
What Systems Are Covered
SB 5 applies to any "automated employment-related decision process" (AERDP) that produces a score, rank, constraint, recommendation, or classification affecting hiring, promotion, discipline, or termination decisions. The definition is broad.
Covered tools include:
- Resume-screening and applicant-tracking systems
- Video-interview analytics and structured interview scoring
- Skills and personality assessments with automated scoring
- Background-screening workflows with automated flags
- Internal promotion and succession-planning engines
- Performance-management tools generating automated ratings
- Scheduling and productivity tools used in discipline or termination decisions
- Reduction-in-force models identifying employees for selection or retention
If a tool generates any output that affects an employment decision and is not a minor factor, it likely qualifies.
Transparency Requirements
Employers must disclose to applicants and employees before using an automated tool that it may be involved in employment decisions. The notice must explain in plain language what the tool does, what personal data it processes, what outputs it generates, and how those outputs may be used.
When an automated system contributes to an adverse decision-rejection, demotion, discipline, or termination-employers must provide a separate disclosure. This statement must explain the principal reasons for the decision, including the degree to which the automated tool contributed, the types of data it processed, and the data sources.
The requirement mirrors similar laws in California and Illinois.
Anti-Bias Testing Is Now Expected
Connecticut amended its Fair Employment Practices Act to make anti-bias review a practical necessity. Courts and enforcement agencies will consider whether employers conducted anti-bias testing and how they responded to findings of disparities.
Employers cannot use "an algorithm made the decision" as a defense. Responsibility for discriminatory outcomes remains with the employer, regardless of whether a vendor or tool contributed.
HR teams should evaluate whether automated tools produce statistically significant disparities based on protected characteristics such as race, gender, or age. Where disparities exist, employers must assess whether the decision is job-related, whether less discriminatory alternatives exist, and what mitigation measures are needed.
Human oversight matters. Decision-makers should understand how tools work, retain authority to override automated outputs, and document their reasoning when they do.
Enforcement and Timeline
Connecticut's attorney general will enforce the employment provisions exclusively under the state's Unfair Trade Practices Act.
Because different provisions take effect at different times, a phased approach works better than waiting for all rules to be final:
- Now through Q3 2026: Inventory all automated tools. Identify which ones qualify as AERDPs. Review vendor agreements. Assign internal owners.
- Q3-Q4 2026: Draft plain-language disclosures. Create internal escalation and human-review workflows. Begin bias testing for high-impact tools.
- Q4 2026-2027: Train HR and managers. Update applicant and employee materials. Refresh procurement standards. Monitor regulatory guidance.
- Ongoing: Repeat audits. Update notices when tools change. Document mitigation efforts. Reassess tools when job criteria or business practices change.
HR leaders should monitor guidance from Connecticut regulators and watch for any enforcement announcements or amendments to the law.
For HR professionals responsible for recruitment technology and talent management decisions, understanding these requirements now prevents costly compliance issues later. AI for Human Resources covers the broader landscape of AI in recruitment automation and talent management systems.
Your membership also unlocks:
