Fired by an Algorithm? What California Law Says When AI Helps Decide Layoffs

Using artificial intelligence to help decide layoffs is not automatically illegal in California, but the technology does not lower the legal bar. If an AI tool or automated system selects people for layoff in a way that discriminates against a protected group, that can violate California's Fair Employment and Housing Act (FEHA). As of October 1, 2025, state regulations make clear that this law applies fully to these tools. “The algorithm decided” is not a defense.

If you were laid off by a process that ran partly or entirely on software, here is what your rights look like in California.

At a glance

• California law does not ban AI in layoff decisions, but it does prohibit outcomes that discriminate against a protected group, whether the bias was intended or not.
• New regulations effective October 1, 2025 confirm that FEHA applies to automated-decision systems, and they extend liability to the outside vendors and agents that build or run the tools.
• Employers must now keep automated-decision records for four years, which can matter a great deal if you need to understand how a layoff was made.
• Blaming the software is not a shield. If an automated layoff harmed you because of your age, race, disability, sex, or another protected trait, it is worth having reviewed.

Is it legal for companies to use AI to decide layoffs?

Yes, using AI or an automated system to help make layoff decisions is generally lawful in California, as long as the outcome does not discriminate. The problem is not the tool itself. The problem is what the tool does. If an automated system ranks, scores, or selects employees in a way that falls more harshly on a protected group, or that targets people because of a protected trait, that result can be unlawful under FEHA even when no human intended to discriminate. California treats the outcome as what matters, not the machinery behind it.

What are California's automated-decision system rules?

California's Civil Rights Council finalized regulations that took effect on October 1, 2025, clarifying how FEHA applies to what the state calls an automated-decision system, or ADS. An ADS is broadly defined as a computational process that makes or helps make a decision about an employment benefit, and it can be built on artificial intelligence, machine learning, algorithms, or statistical models. Resume screeners, ranking and scoring tools, productivity-monitoring systems, and layoff-selection models can all fall within that definition.

The regulations apply to employers with 5 or more employees, the same broad threshold as the rest of FEHA. A few points stand out for employees:

Discrimination through an ADS is unlawful.

It is a violation to use an automated tool that discriminates against applicants or employees based on a protected characteristic, either directly or through a disparate impact, meaning a neutral-looking process that lands harder on a protected group.

Vendors and agents can be on the hook.

Liability extends to an employer's “agents,” which can include the third-party companies and staffing firms that build or operate the tool on the employer's behalf.

Records must be kept for four years.

Employers must preserve automated-decision data, including the inputs, outputs, and selection criteria, for at least four years. That paper trail can be important evidence.

Bias testing cuts both ways.

The rules do not force employers to test their tools for bias, but whether they did, and what they did about the results, can help or hurt them if a discrimination claim is brought.

How can an automated layoff be discriminatory?

The clearest way is disparate impact. Imagine a layoff-scoring model that weighs recent training certifications, tenure patterns, or “flexibility” scores in a way that quietly disadvantages older workers, or a system trained on past decisions that already carried bias. The model can look neutral on its face and still produce a group of laid-off employees that skews heavily toward one protected class. When that happens, the fact that a computer produced the list does not make the result lawful.

Automated tools can also raise other FEHA problems. A system that screens people out based on disability-related information, or that penalizes protected activity such as taking leave, can create liability the same way a biased human decision would. The law looks at who was harmed and why, not at how sophisticated the software was.

Does “the algorithm made the decision” protect my employer?

No. An employer cannot outsource its legal responsibility to a piece of software. FEHA holds the employer accountable for discriminatory outcomes regardless of whether a person or a program produced them, and the 2025 regulations reinforce that by extending responsibility to the vendors and agents involved. If anything, an employer that used an automated tool it never tested for bias may have a harder time defending the result, not an easier one.

What about proposed AI laws that did not pass?

California has considered broader legislation that would have required employers to disclose when AI was used in a decision and to allow appeals, including bills introduced in 2024 and 2025. Those proposals were not enacted, so there is not yet a standalone statute requiring that kind of notice for most private employees. For now, the enforceable protection sits in FEHA and the Civil Rights Council's automated-decision regulations. Separate privacy rules governing automated decision-making technology are also developing, which is one more reason this area is worth watching. Because the law here is moving quickly, it is smart to check the current rules rather than rely on older guidance.

How do I know if my automated layoff was unlawful?

You usually cannot tell from the outside, and we will not pretend a web page can. What tends to matter is whether the layoff fell disproportionately on a protected group, what data and criteria the tool used, whether the employer tested it, and how your situation compares to others who were kept. Much of that information sits in records the employer is now required to preserve for four years. The practical move is to preserve your own documents, note who and what was involved in the decision, and have the facts reviewed before a deadline passes.

What to do if you think AI played a role in your layoff

Take a few steps to protect yourself:

• Write down what you were told about how the layoff decisions were made, including any mention of software, scoring, or a “data-driven” process.
• Save your performance reviews, any metrics or scores you were given, and your layoff paperwork. Keep your own copies and do not access company systems after separation.
• Note the makeup of who was let go versus who stayed, if you know it.
• Talk to an employment lawyer promptly, because discrimination claims carry deadlines and some are short. Do not calendar a date from a guide. Confirm it against your facts.

Frequently asked questions

Is it legal for companies to use AI to decide layoffs?

Generally yes, as long as the outcome does not discriminate. Using an automated tool is lawful, but if it selects people in a way that harms a protected group, intentionally or not, that can violate California's FEHA.

Can I sue if an algorithm was used to lay me off?

Possibly, if the automated decision discriminated against you based on a protected characteristic. California's 2025 regulations confirm FEHA applies to these tools, and liability can extend to the vendor that built them. Whether you have a viable claim depends on the facts, which is worth reviewing.

Does my employer have to tell me AI was used?

There is no broad standalone law requiring that disclosure for most private employees, because bills proposing it were not enacted. But if an automated tool produced a discriminatory result, that can still be unlawful under FEHA regardless of whether you were told.

Can the software company be responsible too?

It can. The 2025 regulations extend liability for automated-decision discrimination to an employer's agents, which can include the third-party vendors and staffing firms that build or run the tool.

How long do I have to take action?

It depends on the claim, and some deadlines are short. Do not rely on a guide for your date. Speak with a lawyer promptly so it is calculated against your actual facts.

Were you laid off by an automated process?

If your layoff ran on software and you believe it harmed you because of your age, race, disability, sex, or another protected trait, it is worth having the facts reviewed before a deadline passes. We represent employees, only employees, across California, and most of our work is in Los Angeles. We handle these matters on a contingency-fee basis, and the first case review is free.

Call (424) 255-8376 or contact us for a free, confidential case review.

The Law Offices of Jonathan J. Delshad is a Los Angeles based employment law firm representing employees across California in wrongful termination, discrimination, retaliation, harassment, and wage and hour matters. Representing employees is the core of the firm's practice. Mr. Delshad serves as Editor-in-Chief of the California Wrongful Termination Law Review and trained at Latham & Watkins. Recognition includes Super Lawyers (2022 to 2025), Best Lawyers (since 2017), and an Avvo 10.0 “Superb” rating. Reviewed for California employment law accuracy. Last updated: July 3, 2026.

Attorney advertising. This article is educational only and is not legal advice. Reading it does not create an attorney-client relationship, which exists only under a signed engagement agreement. Every case is different, and outcomes depend on the specific facts. Deadlines can run early, so consult a lawyer promptly about your situation.