Know what’s on your site before a plaintiff’s firm does
A free privacy audit shows which cookies, pixels and trackers are running — and which ones commonly appear in privacy claims.
For years, companies have announced layoffs using much of the same corporate vocabulary: restructuring, operational efficiencies, organizational realignment and changing market conditions. Artificial intelligence is giving regulators a reason to question what is actually behind those decisions.
If a company eliminates 500 jobs after investing millions of dollars in artificial intelligence, should it be required to disclose the relationship between the two? What if an AI system was trained using the work product of the very employees it is replacing? And what happens when an employer uses employee surveillance data to determine which workers are expendable?
These are no longer hypothetical regulatory issues.
New York, Connecticut and California have taken steps to require employers to disclose when artificial intelligence contributes to certain workforce reductions. The changes build upon a federal employment law that predates the commercial internet: the Worker Adjustment and Retraining Notification (WARN) Act of 1988.
Nearly four decades later, that law is finding new relevance.
For chief privacy officers, general counsel, human resources executives and corporate compliance departments, the consequences extend well beyond employment law. How companies collect employee data, train AI models, conduct internal assessments and document their adoption of artificial intelligence could become matters of regulatory scrutiny and litigation.
What Is the WARN Act?
The Worker Adjustment and Retraining Notification Act is a federal law enacted in 1988 to provide workers and communities with advance notice of qualifying mass layoffs and plant closures.
Under the federal WARN Act, covered employers generally must provide at least 60 calendar days of advance written notice. The law typically applies to employers with 100 or more employees, subject to specific workforce calculations, layoff thresholds and statutory exceptions.
Employers that violate the law may be liable for back pay, benefits and other penalties.
Several states have enacted their own versions of the legislation, commonly referred to as mini-WARN Acts, which in some cases impose more stringent requirements than federal law.
The original legislation was largely conceived around factory closures, manufacturing jobs and corporate restructuring. Congress was not contemplating a world in which an AI agent could assume the responsibilities of an entire customer service department.
The U.S. Department of Labor’s WARN Act guidance outlines existing federal notification requirements. There is not, however, a general federal WARN Act requirement specifically mandating disclosure of AI-related workforce displacement.
That is beginning to change at the state level.
States Are Starting to Require AI Layoff Notifications
New York
In January 2025, New York Gov. Kathy Hochul announced plans to require employers submitting WARN notices to disclose whether layoffs were related to their use of artificial intelligence.
The state subsequently updated its WARN filing portal to request AI-related disclosures from employers.
New York already has more stringent notification requirements than its federal counterpart. Its WARN Act generally applies to private businesses with 50 or more full-time employees and requires 90 days of advance notice for qualifying employment actions.
New York legislators have also introduced the Artificial Intelligence Workforce Impact Transparency Act (S8928), which seeks to formalize and expand AI-related workforce reporting.
As of November 2026, the administrative reporting requirement should be distinguished from proposed statutory changes and amendments to the state’s WARN regulations.
Connecticut
Effective October 1, 2026, Connecticut requires employers submitting federal WARN Act notices to the Connecticut Department of Labor to disclose whether their layoffs relate to the employer’s use of artificial intelligence or another technological change.
The requirement was enacted under the Connecticut Artificial Intelligence Responsibility and Transparency Act, also known as Senate Bill 5 and Public Act 26-15.
The inclusion of technological changes beyond artificial intelligence is particularly noteworthy. Companies cannot necessarily avoid the disclosure requirement simply because they characterize a new system as workflow automation rather than AI.
California
On September 30, 2026, California Gov. Gavin Newsom signed Senate Bill 951, expanding the state’s existing WARN Act to address technological displacement.
The legislation requires additional disclosures when a qualifying mass layoff, relocation or termination is caused entirely or substantially by AI or other automated technology replacing or automating employment positions.
Among the required disclosures is information concerning the job functions that will be automated.
The legislation also requires California’s Employment Development Department to report to the Legislature by January 1, 2028, on AI’s effects on business hiring practices.
For employers preparing for 2027, California’s legislative changes warrant particular attention.
When Is a Layoff Actually an AI Layoff?
This is where the legislation becomes complicated.
Imagine a financial services company that employs 300 customer service representatives. The company introduces an AI-powered customer service platform that can handle 70% of incoming inquiries. Six months later, management eliminates 150 positions.
Was the layoff caused by artificial intelligence, declining operational costs or a corporate restructuring?
All three could be true.
Consider a software company that eliminates 200 engineering positions while increasing its investment in generative AI coding tools. Or an insurance company that reduces its claims processing department after implementing automated document review.
Artificial intelligence does not need to replace an entire employee to contribute to that employee’s displacement. It simply needs to absorb enough of the employee’s responsibilities to change the economics of employing that person.
Regulators will increasingly have to distinguish between businesses adopting AI while independently reducing headcount and businesses reducing headcount because they adopted AI.
Corporate documentation could become central to establishing that distinction.
Internal emails, board presentations, AI procurement contracts, financial projections and workforce planning documents could all become relevant evidence in an employment dispute.
A company telling investors that artificial intelligence will save $50 million in annual labor costs may have difficulty explaining to regulators why subsequent layoffs had nothing to do with automation.
The Privacy Problem: Training AI to Replace Your Own Employees
One of the more uncomfortable questions surrounding workplace AI is whether employees know how their own data is being used.
Consider what an employer has access to:
-
Recorded customer service calls and transcripts
-
Employee emails and internal communications
-
Screen recordings and productivity monitoring data
-
Proprietary code written by software engineers
-
Sales calls and CRM activity
-
Employee performance evaluations
-
Internal documents and knowledge repositories
-
Voice recordings and biometric information
Much of this information can be valuable for training, configuring or improving artificial intelligence systems.
An employee may have been informed that their customer service calls were recorded for quality assurance purposes. They may not have been informed that those recordings could subsequently be used to develop an AI agent capable of performing their job.
From a data privacy perspective, those are two materially different purposes.
Depending on the jurisdiction and circumstances, repurposing employee information in this manner could implicate transparency obligations, purpose limitation, data minimization, lawful processing requirements and employee privacy rights.
The fact that a company owns an internal database does not necessarily mean it has unrestricted rights to process all the personal information contained within it.
California Is Bringing Employment Law and Data Privacy Together
California is particularly interesting because employers face overlapping obligations under employment law and the California Consumer Privacy Act (CCPA).
Since January 1, 2023, employee personal information has generally fallen within the scope of the CCPA for covered businesses.
Beginning January 1, 2027, businesses using covered automated decision-making technology (ADMT) to make significant decisions must also comply with applicable new requirements finalized by the California Privacy Protection Agency.
Depending on the activity and available exceptions, these requirements can include pre-use notices, risk assessments, access rights and mechanisms for opting out of certain automated decision-making activities.
Now consider an employer that implements AI software to analyze workers’ productivity, rank employees based on performance and recommend which employees should be terminated.
That employer could potentially face multiple, overlapping legal obligations involving:
-
California WARN Act notifications
-
CCPA employee privacy rights
-
Automated decision-making disclosures
-
Employment discrimination laws
-
Workplace surveillance restrictions
-
Privacy risk assessments
-
Third-party AI vendor management
A WARN Act notification does not satisfy an employer’s data privacy obligations. Equally, an employee privacy notice does not substitute for a legally required layoff notification.
Businesses will need to address these requirements separately.
Could AI-Related Layoffs Become the Next Wave of Employment Litigation?
It is not difficult to envision a new category of lawsuits involving AI-related workforce displacement.
Employment attorneys may begin investigating whether companies properly disclosed the role of artificial intelligence in layoffs, whether automated systems disproportionately selected employees belonging to protected classes, or whether employers improperly collected and repurposed employee data.
Potential legal disputes could involve allegations of wrongful termination, age discrimination, disability discrimination, unlawful workplace surveillance, privacy violations and failures to provide statutory notifications.
The discovery process alone could become expensive.
Plaintiffs’ attorneys could seek access to AI training datasets, model outputs, vendor agreements, internal prompts, automated employee rankings, management communications and financial models calculating anticipated labor savings.
Companies that have implemented AI without a formal governance program may struggle to reconstruct how decisions were made months or years earlier.
There is also an emerging risk for employment practices liability insurers and cyber insurers. As employment decisions become more dependent on personal data and third-party AI systems, the distinction between a traditional employment claim and a technology-related privacy claim may become less clear.
The European Union Is Already Regulating AI in the Workplace
The United States is not alone in examining the relationship between artificial intelligence and employment.
Under the EU AI Act’s Annex III, certain AI applications used in employment and workforce management fall within the high-risk classification, subject to the Act’s specific classification rules.
These include systems intended to assist with recruitment, employment termination, promotion decisions, employee performance monitoring and certain task allocation decisions.
The General Data Protection Regulation (GDPR) also imposes restrictions on certain solely automated decisions that produce legal or similarly significant effects on individuals under Article 22.
For multinational employers, introducing an AI-powered workforce management system is increasingly a cross-border compliance exercise.
A system that is permissible in one jurisdiction may require additional disclosures, assessments, safeguards or human oversight in another.
The AI Layoffs That Never Get Reported
There is a fundamental limitation to using the WARN Act to measure AI-related job displacement.
What happens when companies simply stop hiring?
A business with 1,000 employees could gradually reduce its workforce to 600 employees over several years through natural attrition, retiring workers, unfilled vacancies and the elimination of entry-level positions.
Depending on the circumstances, the company may never trigger a WARN Act notification.
Yet artificial intelligence could have effectively eliminated 400 employment opportunities.
The same problem exists when companies discontinue contracts with independent contractors, reduce their reliance on outsourced workers or conduct workforce reductions that fall below applicable reporting thresholds.
If governments genuinely want to understand the economic consequences of artificial intelligence, mass layoff notifications will only tell part of the story.
The jobs that never get created may ultimately be as significant as the jobs that disappear.
What Businesses Should Do Before Implementing AI
Companies should not wait until they are preparing mass layoff notifications to assess the legal implications of their AI systems.
A comprehensive AI governance and data privacy program should include:
AI inventory. Maintain a centralized record of artificial intelligence systems used throughout the organization, including AI features embedded within existing enterprise software.
Employee data mapping. Identify what employee information is collected, where it is stored, who has access to it and whether it is used to train or operate AI systems.
AI impact assessments. Evaluate how new technologies could affect employee privacy, discrimination risks, employment decisions and workforce displacement.
Employee privacy notices. Review existing disclosures to determine whether they accurately describe workplace monitoring, automated decision-making and the processing of employee information through AI.
Vendor due diligence. Establish whether third-party AI providers retain employee data, use it for model training or share it with subcontractors.
Human oversight. Establish documented procedures for reviewing AI-assisted employment decisions and challenging inaccurate automated recommendations.
Regulatory change management. Monitor state, federal and international requirements governing AI disclosures, employment notifications and data privacy.
Recordkeeping. Preserve appropriate documentation explaining why AI systems were introduced, how they were assessed and what role they played in workforce decisions, consistent with applicable retention requirements.
These responsibilities should not reside exclusively with human resources or IT. Legal departments, chief privacy officers and executive management should be involved before deployment.
The Next Frontier of Corporate Transparency
For decades, businesses have been required to disclose financial risks, material cybersecurity incidents, data breaches and certain employment decisions.
AI adoption is becoming another area of corporate transparency.
It is not unreasonable for businesses to pursue automation. Technological advancements have displaced jobs throughout history, and companies have legitimate economic reasons to invest in tools that improve productivity.
But there is a meaningful difference between replacing a worker with a machine and building a machine using that worker’s personal information, recorded conversations, behavioral patterns and accumulated professional knowledge.
Should employees have a right to know when their data is being used to automate their own jobs?
Should companies disclose how much of their anticipated AI cost savings will come from workforce reductions?
Should workers be entitled to understand how an algorithm contributed to their termination?
These are questions legislators will increasingly confront.
The WARN Act may have been written in 1988, but its next chapter could be defined by artificial intelligence.
Preparing for the Convergence of AI Governance and Data Privacy
At Captain Compliance, we believe artificial intelligence governance and data privacy should be treated as interconnected corporate responsibilities.
Businesses need visibility into the personal information they collect, the technologies that process it, the third parties that receive it and the legal obligations associated with those activities.
Developing a comprehensive AI governance framework, conducting privacy impact assessments, maintaining accurate privacy notices and managing data subject rights are increasingly important parts of that process.
Companies spent the past several years figuring out how to implement artificial intelligence.
The next several years may be spent explaining how they used it.
Put this into practice
Turn privacy guidance into working controls.
Captain Compliance helps teams discover tracking technologies, enforce consent, manage privacy requests and document the evidence behind every decision.