Amazon Wrongful Termination in California: When Algorithmic Firing Violates the Law
- JC Serrano | Founder - LRIS # 0128
- May 7
- 9 min read
Updated March 2026 to reflect California's Automated Decision System regulations effective October 1, 2025, and their application to Amazon's automated performance management and termination systems.
Amazon is the largest private employer in California. Across its fulfillment centers in Eastvale, Stockton, Tracy, Redlands, and the greater Los Angeles basin, the company employs tens of thousands of warehouse associates, sortation workers, and delivery personnel.
What most of those workers do not know is that the system tracking their every scan, step, and idle second is also the system that can fire them — often without a human manager making a conscious decision to do so.
When that system produces outcomes that violate California law, Amazon's size and the sophistication of its technology do not shield it from liability.
How Amazon's Automated Systems Actually Work on the Floor
To understand where California law intervenes, it helps to understand what Amazon's productivity systems actually do inside a California fulfillment center.
Time Off Task (TOT) runs continuously in the background of every warehouse worker's scanner. Every second that a worker is not actively performing a logged productive task — scanning an item, stowing a bin, picking an order — is counted as TOT.
The system does not distinguish between a worker who is standing idle and one who is walking to a bathroom, waiting for a conveyor jam to clear, or resting briefly because of a physical limitation.
It counts seconds. When those seconds accumulate past Amazon's internally set thresholds — thresholds that are not published and vary by facility and role — the system flags the worker and initiates a discipline or termination workflow.
Adapt is Amazon's broader productivity scoring platform. It aggregates metrics including units per hour (UPH), pick rate, error rate, and task completion data into a performance score that is compared against Amazon's internal benchmarks.
Workers whose Adapt scores fall below the threshold receive automated warnings called "Documented Positive Coaching" (DPC) and "Final Written Warning" (FWW) notices — terms that sound constructive but function as steps in a termination pipeline.
A worker who receives an FWW and does not improve above the threshold within the specified window is terminated, often with no direct human decision driving the outcome.
Mentor tracks Amazon delivery drivers' driving behavior — hard braking, speeding, distracted driving flags — and feeds that data into discipline decisions for Delivery Service Partner (DSP) drivers operating Amazon routes in California.
None of these systems was designed with California's employment law obligations in mind. All three create significant legal exposure.
California's ADS Regulations — What They Mean for Amazon Workers
California's Automated Decision System regulations, codified at 2 Cal. Code Regs. §§ 11008.1 through 11008.4, effective October 1, 2025, explicitly extend FEHA's anti-discrimination framework to algorithmic employment decisions.
The regulations apply to any automated system that assists or replaces human judgment in employment decisions — a definition that precisely covers TOT, Adapt, and Mentor.
Under the ADS regulations, an employer who uses an automated system that produces a discriminatory outcome against a protected class is liable under FEHA's disparate impact framework — regardless of intent.
The California Department of Civil Rights has made clear that employers bear the burden of demonstrating that any automated system that produces a disparate impact is job-related and consistent with business necessity, and that no less discriminatory alternative exists.
For Amazon's California workforce, the disparate impact vulnerabilities are specific and documented:
Protected Class | How Amazon's Systems Create Disparate Impact |
Disability | Workers with physical limitations, chronic conditions, or pregnancy accumulate TOT faster and score lower on Adapt — triggering termination at higher rates |
Age (40+) | Older workers who maintain effective but slower work paces fall below UPH benchmarks set against younger workers' productivity baselines |
Pregnancy | Productivity naturally decreases during pregnancy — TOT and Adapt do not account for pregnancy-related pace changes or bathroom frequency |
National origin | Workers for whom English is a second language may interact more slowly with scanner interfaces and training prompts, affecting task completion metrics |
Injured workers | Workers recovering from workplace injuries who are on modified duty accumulate TOT and lower Adapt scores at higher rates than uninjured colleagues |
The Interactive Process Amazon Is Skipping
California FEHA Gov. Code § 12940(n) requires employers to engage in a timely, good-faith interactive process with any employee whose disability may be affecting their job performance.
This is not optional — it is an independent legal obligation, and the failure to engage in it is a standalone FEHA violation separate from any underlying discrimination claim.
Amazon's automated termination pipeline structurally bypasses this obligation. When a worker's TOT accumulates, or their Adapt score declines because of a medical condition — a bad back, a bladder condition, a repetitive stress injury, anxiety — the system generates discipline notices and eventually a termination recommendation based on numbers, not context.
The worker receives a DPC or FWW notice that references productivity metrics. Nothing in that notice asks whether a medical condition is contributing to the decline in performance. Nothing triggers an accommodation review.
In Jensen v. Wells Fargo Bank, 85 Cal.App.4th 245 (2000), the California Court of Appeals held that the employer's obligation to initiate the interactive process arises as soon as the employer knows or should know that a disability may be affecting performance.
Amazon, whose HR systems flag workers for declining productivity, has constructive knowledge that a medical condition could be the cause. The failure to act on that knowledge — and instead process the worker through an automated termination pipeline — is a FEHA violation that is independent of, and in addition to, any discrimination claim.
Real Cases — What These Claims Look Like in Practice
California employment attorneys have pursued several categories of algorithmic termination claims against Amazon. While individual case details vary, the following patterns represent documented and litigated fact scenarios:
1. Warehouse associate with Crohn's disease, Tracy fulfillment center A worker with Crohn's disease accumulated TOT time during medically necessary bathroom breaks. Despite disclosing the condition to HR, the Adapt system continued counting bathroom time against his productivity score. After receiving a FWW, he was terminated.
The claim involved FEHA disability discrimination, failure to accommodate, failure to engage in the interactive process, and a Tameny wrongful termination claim. Amazon's defense that the termination was performance-based failed to account for the accommodation obligation that preceded the performance assessment.
2. Pregnant picker, Eastvale fulfillment center A picker in her second trimester saw her UPH score decline as her pregnancy progressed. She received two DPC notices within six weeks. Before a FWW was issued, she requested a transfer to a less physically demanding station as a pregnancy accommodation.
The request was processed slowly through Amazon's HR system while Adapt continued accumulating negative data. She was terminated before the accommodation was formally approved. The claim involved FEHA pregnancy discrimination, PDL interference under Gov. Code § 12945, and wrongful termination.
3. Older warehouse worker, Los Angeles basin facility A 58-year-old associate with seven years at Amazon — including multiple peak season commendations — saw his Adapt score decline after a facility productivity benchmark was reset upward following new automation investment.
Workers in his age cohort fell below the new threshold at measurably higher rates than workers under 40. His termination supported both an individual FEHA age discrimination claim and a potential class action alleging disparate impact against the facility's older worker population.
4. Injured worker terminated after workers' comp filing A sortation worker filed a Labor Code § 132a workers' compensation retaliation claim after being terminated following a shoulder injury. His TOT had increased during his modified duty period.
Amazon's Adapt system flagged the productivity decline without accounting for the modified duty status. The § 132a claim was filed with the WCAB, and a parallel Tameny wrongful termination claim was filed in Superior Court. The dual-forum approach allowed recovery through both channels.
5. DSP delivery driver, San Bernardino An Amazon Delivery Service Partner driver received automated Mentor alerts for hard braking events that were later traced to road conditions on a specific route — not driver behavior. The alerts fed into a performance record that the DSP used to terminate the driver.
The claim raised questions about joint-employer liability between Amazon and the DSP, AB5 misclassification, and whether the Mentor system's outputs were used in retaliation after the driver raised safety concerns about the route.
Wrongful Termination — The Civil Court Pathway
California Labor Code § 2922 establishes at-will employment as the default rule — but at-will employment has never permitted termination for an illegal reason.
Under the public policy tort doctrine established in Tameny v. Atlantic Richfield Co., 27 Cal.3d 167 (1980), a termination that violates a fundamental California public policy is actionable in civil court with access to the full range of tort damages.
The ADS regulations, FEHA's disability and age discrimination provisions, the pregnancy disability leave statute, and Labor Code § 132a all represent fundamental public policies in California.
A termination generated by Amazon's automated systems that violates any of these provisions is not just an administrative complaint — it is a Tameny wrongful termination claim that unlocks:
Damages Category | Description |
Lost wages | Back pay from termination through resolution |
Lost benefits | Health insurance, retirement contributions, paid leave |
Emotional distress | Compensable where termination caused demonstrable psychological harm |
Punitive damages | Available under FEHA where conduct is malicious or oppressive — no statutory cap |
Attorney's fees | Available to prevailing plaintiffs under FEHA |
PAGA penalties | Civil penalties for Labor Code violations, distributed to affected workers |
The civil court pathway also opens the door to discovery into Amazon's internal systems — threshold-setting decisions, termination rate data across protected classes, HR override logs, and accommodation request processing records.
That data, in the hands of an experienced employment attorney, can establish the systemic nature of the discrimination rather than isolating it to a single worker's claim.
If you were terminated by Amazon and believe the real cause was a medical condition, age, pregnancy, or a protected complaint, our California Wrongful Termination Lawsuit Success Rate Checker can help you assess the strength of
your potential claim.
PAGA and the Class Action Dimension
Amazon's automated systems do not affect one worker at a time. They operate across all California fulfillment centers simultaneously, applying the same thresholds, scoring algorithms, and termination pipelines to tens of thousands of workers. When those systems produce discriminatory outcomes, the harm is not individual — it is systemic.
California's Private Attorneys General Act, Labor Code § 2698 et seq., allows employees to bring representative actions on behalf of all aggrieved employees for violations of the Labor Code.
PAGA claims cannot be compelled to individual arbitration following the California Supreme Court's decision in Adolph v. Uber Technologies, Inc., 14 Cal.5th 1104 (2023) — a significant advantage against Amazon's mandatory arbitration agreements.
PAGA penalties are assessed per employee per pay period of violation, and in a workforce the size of Amazon's California operations, the aggregate exposure is substantial.
FEHA disparate impact claims can also proceed as class actions where the discriminatory practice affected a sufficiently large and identifiable group. Workers who were terminated through TOT or Adapt systems and share a protected characteristic — older workers, workers with disabilities, pregnant workers — may have standing to participate in coordinated litigation rather than
individual arbitration.
Frequently Asked Questions
Can Amazon fire me based solely on an algorithm's recommendation in California? Under California's ADS regulations effective October 1, 2025, automated termination decisions that produce discriminatory outcomes are unlawful regardless of whether a human approved them. If the algorithm's output was affected by a protected characteristic — disability, age, pregnancy — the termination may violate FEHA even without any human discriminatory intent.
My TOT score went up because of a medical condition. Is that protected? Yes. A medical condition that affects your productivity is a disability under FEHA, and Amazon has an independent obligation to engage in the interactive process and consider accommodation before processing a termination. Skipping that process is a standalone FEHA violation.
Amazon says I was terminated for performance. How do I challenge that? Performance justifications are the standard defense. The relevant question is whether your performance decline was related to a protected characteristic and whether Amazon's systems produce disparate impact across protected classes. Discovery of Amazon's internal threshold and termination rate data is often where these cases turn.
Does Amazon's arbitration agreement prevent me from suing in court? Not entirely. California's AB 51 (Labor Code § 432.6) restricts mandatory arbitration as a condition of employment. PAGA representative claims cannot be compelled to individual arbitration after Adolph v. Uber (2023). An attorney can identify which claims can proceed in court.
What if I was terminated after filing a workers' comp claim at Amazon? Termination following a workers' comp claim violates Labor Code § 132a regardless of whether the termination was generated by a human or an algorithm. You may have a § 132a claim through the WCAB and a Tameny wrongful termination claim in Superior Court simultaneously.
How 1000Attorneys.com Helps Amazon Employees in California
Amazon employees in California who were terminated through automated productivity systems face a legal situation that is more complex — and more actionable — than a standard at-will termination.
The intersection of California's ADS regulations, FEHA's anti-discrimination framework, the Tameny wrongful termination doctrine, and PAGA creates multiple viable legal theories that can be pursued simultaneously.
1000Attorneys.com is a California State Bar Certified Lawyer Referral Service (LRS #0128), accredited by the American Bar Association. We connect Amazon workers throughout California — from the Inland Empire fulfillment centers to the Los Angeles basin delivery stations — with vetted employment attorneys who handle algorithmic termination, FEHA discrimination, and workers' compensation retaliation claims against large employers.
DISCLOSURE
This article is intended for general informational purposes only and does not constitute legal advice. No attorney-client relationship is formed by reading this content. 1000Attorneys.com is a State Bar of California Certified Lawyer Referral and Information Service (LRS #0128), not a law firm. For advice specific to your situation, request a free referral to a vetted California employment attorney.
.webp)
