Connecticut AI WARN Act employer requirements for layoffs and workforce decisions
Connecticut has tied artificial intelligence use directly to layoff notifications under its updated WARN framework, creating a new compliance frontier for chief human resources officers. In 2024, the legislature enacted Public Act 24-5 (often referred to as the Connecticut AI and WARN Act), with core provisions scheduled to take effect on October 1, 2024. When an employer issues a mass layoff or plant closing notice, covered Connecticut employers must now state whether automated tools, algorithmic systems, or decision technology were a substantial factor in the employment decision. This explicit link between AI and workforce reductions forces employers to map where automated employment systems influence headcount, restructuring, or role elimination, and to align those disclosures with the statutory text of Public Act 24-5 and any implementing guidance from the Connecticut Department of Labor.
Under Connecticut law, the WARN notice sent to the state Department of Labor and to affected employees must provide a clear written statement explaining if artificial intelligence or other automated tools meaningfully altered the outcome of the decision. For October employers planning restructurings, this means HR, Legal, and IT must align on which decision technology or AEDT systems contributed to making employment choices and how those systems were used. The new AI-related WARN disclosure duties therefore extend beyond classic employment law notifications and push employers to document AI’s role in every relevant employment decision, including which models were consulted, what data they relied on, and how their outputs were weighed against human judgment. Employers should periodically confirm that their internal summaries match the current text of Public Act 24-5 and any formal Department of Labor FAQs or bulletins.
For a CHRO, the operational impact is immediate and concrete, because every WARN written notice now doubles as an AI governance artifact. You will need an inventory of automated employment tools, including frontier developers’ platforms, and a record of where those tools were a substantial factor in workforce decisions. That inventory must connect to specific employment decisions so that any WARN notice can accurately describe the technology, the purpose, and the categories of personal data processed without creating new discrimination or anti discrimination risks. A simple checklist can help: list each AI or automated decision tool, identify the roles or locations it affects, record when it is used in layoff or redeployment scenarios, and retain sample WARN language that explains the tool’s function in plain, non technical terms. For example, a notice might state that an automated workforce analytics system was used to generate cost reduction scenarios, that managers reviewed those scenarios against established, non discriminatory criteria, and that final selections were made by human decision makers.
Notice, transparency, and bias testing obligations under Connecticut AI law
Beyond WARN, Connecticut law introduces broader transparency duties for automated employment decision technology that meaningfully alters hiring, promotion, or termination outcomes. Public Act 24-5 requires that, before using such technology on any job applicant or current employee, an employer must provide a written notice explaining the purpose of the tools, the types of personal data used, and how the automated employment system will influence the employment decision. These AI-related notice and disclosure obligations apply whether decisions are fully automated or whether AI is only one substantial factor in human led employment decisions, and they sit alongside existing notice requirements under state labor and anti discrimination statutes, as well as any future interpretive guidance from the Connecticut Attorney General or Department of Labor.
The statute expects employers to implement bias testing and anti bias safeguards for any AEDT or decision technology used in making employment choices. Bias testing must evaluate whether automated systems create discrimination risks across protected groups, and Connecticut employers must adjust or suspend tools that show problematic patterns. For HR leaders managing sensitive cases such as harassment allegations, these transparency and fairness duties sit alongside existing rights frameworks, which are explored in depth in this guide on employee rights when accused of harassment. In practice, this means scheduling periodic audits of AI driven tools, documenting test methodologies, and preserving reports that show how disparate impact, false positive rates, and other fairness metrics were evaluated, so that employers can demonstrate compliance if the Attorney General investigates potential violations under the Connecticut Unfair Trade Practices Act.
Enforcement of these AI and employment law provisions rests with the Connecticut Attorney General under the Connecticut Unfair Trade Practices Act, not with private plaintiffs. That structure means the Attorney General can investigate patterns of non compliance, especially where automated tools or artificial intelligence appear to drive systemic discrimination or anti discrimination failures. For CHROs, this raises the bar on documentation, because regulators will expect clear evidence of bias testing, risk assessments, and governance controls around every significant employment decision influenced by AI. Employers should monitor any formal guidance, FAQs, or advisory opinions issued by the Attorney General or the Department of Labor, and keep internal files that show how those materials were translated into updated policies, training, and system configurations, including documented timelines for implementing required changes.
Strategic playbook for CHROs: governance, documentation, and cross state risk
Connecticut’s approach signals a new national paradigm where AI use in employment decisions is treated as a regulated risk, not just a technology choice. Even if your organisation has no operations in Connecticut, the Connecticut AI WARN Act employer requirements preview how other states such as Illinois, Colorado, and Texas may link AI, employment law, and layoff accountability. For CHROs on the path to the C suite, this is a governance test of how well you can align people strategy, decision technology, and legal risk management, and how effectively you can brief the board on emerging AI compliance expectations, including the specific obligations contained in Public Act 24-5 and any comparable statutes in other jurisdictions.
A practical first step is to build a cross functional AI in employment register that lists all automated tools, their frontier developers, their decision logic, and the employment decisions they influence. That register should track where artificial intelligence is a substantial factor in making employment choices, what personal data is processed, and what bias testing or anti bias controls are in place. When you vet external partners or consultants who deploy automated employment systems, use a structured approach such as the one outlined in this analysis of effective ways to vet third party HR consultants. As you refine this register, include sample WARN notice language that explains, for example, that an automated workforce planning tool was used to model cost scenarios but that final layoff selections were made by human managers applying established, non discriminatory criteria, and assign clear ownership for drafting, legal review, and annual updates.
CHROs should also align AI governance with broader fairness frameworks, including how favoritism, discrimination, and perceived unfairness are handled in promotion and layoff decisions. A useful complement is this perspective on whether favoritism in the workplace is illegal, which helps frame how automated systems can amplify or mitigate human bias. As Connecticut employers adapt to written notice duties, Department of Labor reporting, and scrutiny from the Attorney General, HR leaders elsewhere can treat these rules as a blueprint for documenting AI’s role in every significant employment decision before similar WARN style obligations arrive in their own states. A concise CHRO checklist might include: maintaining an up to date AI tool inventory, mapping each system to specific employment decisions, preserving bias testing records, and pre drafting WARN templates that clearly disclose when automated decision technology was a substantial factor in workforce actions, with explicit timelines, accountable owners, and periodic board level reporting.