Connecticut's New AI Law Mandates Employer Accountability and Developer Transparency in AI-Driven Employment Decisions

On May 27, 2026, Connecticut Governor Ned Lamont signed Senate Bill 5 (Public Act No. 26-15) into law, introducing sweeping regulations for the use of artificial intelligence in employment-related decisions. The law, which takes effect on October 1, 2026, for several key provisions, establishes new obligations for both employers utilizing AI in their hiring, promotion, and termination processes, and for the developers creating these AI tools. Notably, SB 5 amends the Connecticut Fair Employment Practices Act (CFEPA) to explicitly state that the use of an Automated Employment-related Decision Technology (AEDT) cannot serve as a defense against discrimination claims. Furthermore, employers issuing layoff notices under the federal Worker Adjustment and Retraining Notification (WARN) Act must now disclose whether these layoffs are related to the employer's use of AI or other technological changes. This legislation is critical for practitioners in HR, legal, and technology roles, as it directly impacts the operational deployment of AI in sensitive employment contexts. The inability to use AI as a defense against discrimination claims means organizations must ensure their AI systems are not only technically sound but also demonstrably fair and unbiased. The disclosure requirement for AI-related layoffs adds a new layer of transparency, potentially influencing public perception and employee relations. For AI developers, the law necessitates a proactive approach to embedding ethical considerations and compliance features into their tools from the outset, rather than as an afterthought. Failure to comply could result in significant civil penalties, enforced by the Attorney General. The Connecticut law fits into a broader, accelerating trend of AI governance and regulation across various jurisdictions. States like New York and California have also been at the forefront of establishing AI-related legislation, particularly concerning employment and data transparency. For instance, New York has previously proposed similar requirements for layoff disclosures, and California has implemented its own AI transparency and frontier AI safety laws. The EU AI Act, with its phased enforcement, also sets a global precedent for comprehensive AI regulation, pushing for transparency and risk management across sectors. This growing regulatory landscape underscores a global recognition that while AI offers immense potential, its deployment must be balanced with robust safeguards to protect individuals and ensure accountability. The NIST AI Risk Management Framework (AI RMF) and ISO 42001 are also becoming de facto standards for organizations seeking to demonstrate responsible AI practices. In practice, organizations using or developing AEDTs in Connecticut should immediately conduct a thorough audit of their AI systems to identify potential areas of non-compliance. This includes reviewing algorithms for bias, ensuring clear documentation of AI's role in decision-making, and establishing transparent communication protocols for employees and applicants. Practitioners should also consider implementing robust human oversight mechanisms for AI-driven employment decisions and invest in training to ensure HR personnel understand the capabilities and limitations of these tools. For developers, this means building in explainability features, bias detection, and auditable logs into their AI products. The evolving regulatory environment demands a shift from simply deploying AI to strategically governing its use, making compliance an integral part of the AI lifecycle.