AI and Layoffs: What Employers Need to Know About AI Disclosure Requirements

Employers are increasingly using AI tools at various points in a reduction in force: workforce analytics platforms to identify which roles are redundant, performance management systems that produce AI-generated ratings used as selection criteria, and data analysis tools that model the demographic impact of proposed selection lists. These tools can be legitimate and useful. They also create disclosure obligations that many employers have not yet mapped to their RIF compliance processes.

The regulatory landscape is developing faster than most HR teams realize. This article covers where the law currently stands, what obligations apply today, and what to watch as this area continues to evolve.

The Federal Landscape

There is no federal law that specifically requires disclosure when AI tools are used in layoff selection decisions. The federal WARN Act does not mention AI and has not been amended to address it. Title VII's disparate impact framework applies to the outcomes of AI-assisted selection the same way it applies to any other selection process: the question is whether the selection produced discriminatory results, not how the selection was made.

What the federal framework does require is that employers be able to document and defend their selection criteria. If an AI tool was used to generate performance scores, identify redundant roles, or rank employees for selection, and those outputs were used in the selection decision, the employer needs to be able to explain and defend the methodology behind those outputs in the same way they would defend any other selection criterion.

The Equal Employment Opportunity Commission has issued guidance indicating that employers cannot insulate themselves from discrimination liability by attributing a selection decision to an algorithm. If the algorithm's output was used to make or inform the selection, the employer is responsible for the outcome. This is not a new legal principle; it is an application of existing disparate impact law to a new type of selection tool.

State AI Disclosure Laws That Apply Now

Several states have enacted laws that require employers to disclose when automated decision tools are used in employment decisions. These laws vary in scope, but the trend is toward requiring notice to employees and documentation of the methodology.

Illinois. The Illinois Artificial Intelligence Video Interview Act (2020) requires employers using AI to analyze video interviews for employment decisions to notify candidates, obtain consent, and explain how the AI is used. While focused on video interviews, the law signals Illinois's appetite for AI transparency regulation in employment contexts.

New York City. Local Law 144 (effective July 2023) requires employers using automated employment decision tools in hiring or promotion decisions affecting NYC employees to conduct an annual bias audit by an independent third party, publish the audit results, and notify candidates and employees that an automated tool is being used. The law currently applies to hiring and promotion decisions; its application to layoff selection has not been directly tested, but the notice and audit requirements are relevant to any employer using automated tools in employment decisions affecting NYC employees.

Colorado. SB 24-205 (effective February 2026) requires developers and deployers of AI systems used in "consequential decisions" to use reasonable care to protect against algorithmic discrimination. Employment decisions, including termination, are explicitly covered as consequential decisions. Employers deploying AI systems in layoff selection that affects Colorado employees are within the scope of this law.

Maryland and other states. Several additional states have proposed or are developing similar legislation. The direction is consistent: transparency about when AI tools are used in employment decisions, documentation of the methodology, and some form of bias auditing or review requirement.

The EU AI Act

For employers with operations in the European Union, the EU AI Act (fully applicable from August 2026) classifies AI systems used to make decisions about employment, including termination, as high-risk. High-risk AI systems are subject to requirements including: conformity assessment before deployment, technical documentation of the system's design and training data, human oversight requirements, transparency to affected individuals, and registration in a public database.

Employers using AI tools to assist in layoff selection affecting EU employees need to assess whether those tools qualify as high-risk AI systems under the Act and ensure the required documentation and oversight processes are in place.

The Adverse Impact Analysis Obligation Does Not Change

Whether or not a specific AI disclosure law applies to your jurisdiction and your specific use of AI tools, the adverse impact analysis obligation is unchanged. Any selection process that results in disparate impact on a protected class is potentially discriminatory, regardless of whether the selection was made by a human, an algorithm, or a combination of both.

The practical implication for AI-assisted selection is that the adverse impact analysis needs to assess the demographic outcomes of the AI-generated outputs, not just the final selection list. If an AI tool was used to generate performance rankings that were then used as selection criteria, and those rankings produced disparate impact on a protected class, the question is whether the AI methodology was job-related and consistent with business necessity, the same standard that applies to any performance evaluation system.

Documenting the methodology behind any AI tool used in selection, and including that documentation in the audit trail of the reduction, is both good practice and an increasingly explicit legal requirement.

What to Do Before Your Next Reduction

If your organization uses any of the following in the context of a RIF, review your disclosure and documentation obligations before proceeding:

• Workforce analytics platforms that identify roles for elimination based on performance, productivity, or skill data
• Performance management systems that generate AI-assisted ratings used as selection criteria
• Demographic modeling tools used to assess adverse impact
• Any third-party tool that takes employee data as input and produces outputs used in selection decisions

For each tool, determine: what data was used as input, what the output was used for, whether the tool's methodology can be documented and explained, and whether any applicable state law requires disclosure to affected employees.

Include that documentation in the audit trail of the reduction. If the selection is later challenged, the employer's ability to produce the methodology behind any AI-assisted output is part of the legal defense, just as the ability to produce the performance evaluation methodology would be for a traditionally scored selection.

The disclosure and documentation requirements in this area are evolving quickly. Employers conducting reductions in states with active AI legislation should confirm the current requirements with outside employment counsel before finalizing the selection process. For the full compliance checklist that applies to any RIF regardless of AI involvement, see the RIF compliance checklist and the RIF compliance checklist for HR leaders.