Skip to main content
NOPE
Medium reported severity

US: 26 Meta employees sue alleging an AI-assisted May 2026 layoff selection penalized workers on protected leave, Meta says people made the decisions

On 20 May 2026 Meta began notifying about 8,000 employees (roughly 10 percent of its workforce) that they had been selected for layoff. On 13 July 2026 twenty-six anonymous employees who had taken or requested medical, pregnancy, parental or family leave, or a disability accommodation, sued Meta in the US District Court for the Northern District of California. The complaint alleges, on information and belief, that Meta used internal AI-assisted systems (including the Metamate assistant, keystroke and activity monitoring, AI-token-usage dashboards and algorithmic performance ranking) to score, rank and select employees, and that these inputs could not accumulate during protected leave, so plaintiffs on leave were disproportionately selected. Meta says workforce decisions were made by people, not AI, and that no selection decision was made by AI. On 17 July 2026 the court denied a temporary restraining order, recording that the parties dispute whether Meta used AI in the terminations and finding serious questions on the merits but no shown likelihood of success. One plaintiff was voluntarily dismissed on 3 August 2026. The preliminary injunction motion was argued on 24 August 2026 and taken under submission. The plaintiffs are pursuing their merits claims in arbitration and the allegations are unproven.

AI system
Meta internal AI systems (reported)
Meta Platforms, Inc.
Occurred
20 May 2026
Reported
14 July 2026
Event location
United States
What the AI did
Made a decision about the person · Made a claim about the person
Reported harm
Professional Harm
Whose AI use
An institution’s AI use
Setting
Work · Health
Evidence
AI involvement disputed · Causal attribution disputed · 10 sources, 5 underlying accounts
9 claims: 3 documented, 6 reported. 8 open questions
People reported harmed
At least 25 people

AI system as recorded: Meta internal AI-assisted systems alleged in the complaint: the Metamate internal LLM assistant, employee-trained "second brain" agents, keystroke and activity-monitoring productivity scoring, AI-token-usage dashboards and AI-assisted performance calibration (Meta denies AI was used in selecting employees for the layoff)

Reported harm

The complaint alleges that 26 employees were selected for termination in Meta’s May 2026 reduction in force through an AI-assisted process that penalised protected leave, with separations scheduled from 22 July 2026 and asserted losses of employer-subsidised health coverage, unvested equity and, for four plaintiffs, work-visa status. Meta denies that AI was used in the selections. The court found the alleged harms other than immigration status remediable in arbitration and has not decided the merits.

What remains unknown

  • Whether any AI system scored, ranked or selected any plaintiff is unresolved. The complaint pleads it on information and belief. Meta denies it in declarations. No discovery had occurred in the sources inspected, and the judge said on 17 July 2026 that plaintiffs were not in the rooms where the selections happened.
  • The sources inspected do not show how many people beyond the plaintiffs were selected because of leave, or how many of the roughly 8,000 laid-off employees were scored by any AI-assisted system. Only the plaintiffs are counted as reported harmed.
  • The order on the preliminary injunction motion (argued 24 August 2026, taken under submission) had not been found by 29 September 2026. The hearing remarks were read only through one news report and the transcript was not inspected.
  • Whether each plaintiff’s separation took place on its scheduled date (22 July 2026 for most, later for some) is not stated in the sources inspected. Reuters and CBS/AP describe the plaintiffs as still employed in July, and Courthouse News called them former employees on 24 August 2026.
  • One plaintiff (Doe 20) was voluntarily dismissed on 3 August 2026 and no reason is stated. The affected count therefore uses 25.
  • The complaint and declarations contain individual medical, pregnancy and immigration details. They are anonymised in the public copies and are deliberately not described here.
  • Reuters and WSJ pages for the initial filing could not be read (HTTP 401). Reuters coverage of the 17 July ruling was read through a Spokesman-Review copy. The Law.com article was behind a subscription wall and was not cited.
  • The complaint alleges that data captured by a separate Meta employee-monitoring program (the Model Capability Initiative) supplied inputs to the selection process, and Meta denies that AI was used in the selection. That program’s own privacy and security consequences are a different event, recorded as 2026-meta-mci-us-employee-keystroke-and-screen-capture-for-ai-training-then-internal-data-exposure, and are not part of this record.

What the evidence supports

AI involvement: disputed. The complaint alleges, on information and belief, that Meta used internal AI-assisted systems (Metamate, second-brain agents, activity-monitoring scoring, AI-token-usage dashboards, algorithmic ranking and calibration) to score, rank and select employees for the 20 May 2026 termination list. Meta’s declarations say selection decisions were made by human business leaders and that there was no AI-assisted scoring or ranking related to employee performance, and the court described the declarations as saying Meta did not use AI in any way with respect to the RIF. The court recorded that the parties dispute whether Meta used AI in deciding to terminate the plaintiffs and found on 17 July 2026 that the plaintiffs had not shown a likelihood of success on the current record. On 24 August 2026 the judge said the plaintiffs’ evidence raised some potential questions about Meta’s categorical denial. No discovery had tested Meta’s account in the sources inspected.

9 claims: 3 documented, 6 reported. What the statuses mean

Documented On 13 July 2026 twenty-six anonymous plaintiffs (Does 1 through 26) filed a complaint against Meta Platforms, Inc. in the US District Court for the Northern District of California, Oakland division, alleging employment discrimination in the May 2026 reduction in force.

Causal attribution. The filing of the complaint is documented by the court record. The complaint establishes only the plaintiffs’ allegations.

Reported The complaint alleges, on information and belief, that Meta used a set of internal AI-assisted systems to score, rank and select employees for the termination list: the Metamate internal large-language-model assistant, employee-trained "second brain" agents, keystroke and other activity-monitoring data, AI-token-usage dashboards, and algorithmically assisted performance ranking and calibration.

Causal attribution. These are the plaintiffs’ allegations, pleaded on information and belief and supported by plaintiff declarations that were not inspected (they are quoted in the complaint and the order). Meta disputes them (claim c4). The news reports and the court order restate the complaint and are one reporting chain with it.

Reported The complaint alleges that the inputs to these systems cannot be accumulated by an employee on protected medical, family or parental leave, that Meta did not neutralise them for protected leave, and that employees who took protected leave were therefore disproportionately selected for layoff. It states that each plaintiff was selected for separation in the reduction in force and had, within the 24 months before it, taken, requested or been approved for protected leave, or requested or received a disability accommodation.

Causal attribution. Allegation of the plaintiffs. The court found on 17 July 2026 only serious questions going to the merits and no likelihood of success on that record (claim c5). Whether any AI system affected any plaintiff’s selection is unresolved.

Reported Meta denies the allegations. A Meta spokesperson said workforce management and organisational decisions were made by people, not AI. In declarations filed by a Meta human-resources director, Meta states that selection decisions were made by human business leaders using documented neutral criteria, that no AI-assisted scoring or ranking of employee performance was used, that no plaintiff was selected on the basis of leave status or disability, and that no selection decision was made by AI.

Causal attribution. Meta’s statements and declarations about its own process. They are a party account and had been tested only through briefing and court hearings (claims c5 and c7), not in discovery.

Documented On 17 July 2026 the court denied the plaintiffs’ motion for a temporary restraining order. The order records that the parties dispute whether Meta used AI in deciding to terminate the plaintiffs, finds serious questions going to the merits but no shown likelihood of success on that record, finds that loss of health care, employment, protected leave and unvested equity can be remedied in arbitration, and treats the possible loss of immigration status for four plaintiffs with Meta-sponsored visas as a potentially irreparable harm that needed more information.

Causal attribution. A court order establishes what the court ruled and recorded. It does not establish that AI was or was not used in the selection.

Reported Meta began notifying employees selected in the reduction in force on 20 May 2026 (about 8,000 employees, roughly 10 percent of its workforce). The complaint says the plaintiffs were notified of their selection from that date, and that separations were scheduled for about 22 July 2026 (later for some plaintiffs). Meta said in court filings that laid-off workers stayed on the payroll but lost access to Meta systems on 20 May.

Causal attribution. The notification date and the scheduled separation dates come from the complaint and Meta court filings as relayed by Reuters. The sources inspected do not report whether each separation took place on its scheduled date, although Courthouse News called the plaintiffs former employees on 24 August 2026.

Reported The court held a hearing on the plaintiffs’ preliminary injunction motion on 24 August 2026, heard argument and took the motion under submission with a written order to follow. Courthouse News reports that the judge said the record at that time did not persuade the judge of the merits, that the plaintiffs’ evidence raised some potential questions about Meta’s categorical denial of any impact of AI in the termination process, and that the case turns on whether what plaintiffs think happened, happened.

Causal attribution. The minute entry documents that the hearing occurred and the motion was taken under submission. The judge’s remarks are reported by one news outlet and the transcript was not inspected. No written ruling on the preliminary injunction was found as of 29 September 2026.

Documented One plaintiff (Doe 20) voluntarily dismissed all of that plaintiff’s claims without prejudice on 3 August 2026 (Dkt. 33), and the notice states that Does 1 through 19 and 21 through 26 remain parties. No reason is given in the notice.

Causal attribution. A filing by the plaintiffs’ counsel records the dismissal. It does not say whether that plaintiff was selected in the reduction in force or why the dismissal was made.

Reported The 17 July 2026 order records that the plaintiffs recognize their claims must be arbitrated and that they have initiated arbitration proceedings with the American Arbitration Association. The complaint says the plaintiffs seek from the court only status-quo-preserving provisional relief pending an independent audit of the selection process and arbitration of the merits, and it reserves merits relief such as reinstatement, back pay and damages to the arbitral forum.

Causal attribution. The order records the plaintiffs’ own statement about their arbitration filings. The arbitration itself is confidential and its progress is unknown.

Sources

10 sources inspected, from 5 underlying accounts. Sources that repeat one account do not corroborate each other.

How the sources were read, and where the events happened

Event countries: United States. Affected-person countries: United States. Court countries: United States.

The complaint states that the plaintiffs worked in California, Illinois, Washington, New York, the District of Columbia, Pennsylvania and Florida and the suit is in the US District Court for the Northern District of California. Meta also notified employees in other countries in the same reduction in force, but no plaintiff outside the US is described.

Reviewed for publication 2026-09-29: The complaint, the court order denying the temporary restraining order, two Meta declarations, the plaintiffs’ notice of dismissal of one plaintiff and the hearing minute entry were read from the court record (CourtListener copies), with CBS/AP, Reuters and Courthouse News reports. Every AI-related statement is attributed. The plaintiffs are anonymous in the record and are not described individually. Meta denies that AI made or informed any selection decision and the court has not resolved the dispute.

People reported harmed in this case

At least 25 people

0 AI participants · 25 other people harmed

The complaint states that each of 26 plaintiffs was selected for separation in the May 2026 reduction in force. One plaintiff was voluntarily dismissed on 3 August 2026 without a stated reason, so 25 active plaintiffs are counted as a documented minimum. The roughly 8,000 other employees notified in the reduction in force are not counted because no source ties their selection to the alleged AI process.

Counted once within this case. The same person may appear in other cases. This count does not establish AI causation.

Cite this case

Compiled per our published methodology: verification statuses, sourcing standards, and corrections process.

APA

NOPE. (2026). US: 26 Meta employees sue alleging an AI-assisted May 2026 layoff selection penalized workers on protected leave, Meta says people made the decisions. AI incidents. https://nope.net/incidents/2026-us-meta-26-employees-sue-alleging-ai-assisted-may-layoff-selection-penalized-protected-leave

BibTeX

@misc{2026_us_meta_26_employees_sue_alleging_ai_assisted_may_layoff_selection_penalized_protected_leave,
  title = {US: 26 Meta employees sue alleging an AI-assisted May 2026 layoff selection penalized workers on protected leave, Meta says people made the decisions},
  author = {NOPE},
  year = {2026},
  howpublished = {AI incidents},
  url = {https://nope.net/incidents/2026-us-meta-26-employees-sue-alleging-ai-assisted-may-layoff-selection-penalized-protected-leave}
}

Related cases

Medium GPT-5.4 and Claude Opus 5 (reported)

Massachusetts: fired teacher asks court to vacate an arbitration, saying the arbitrator improperly used AI tools to judge her AI-detector evidence

A kindergarten teacher at a public elementary school in Concord, Massachusetts, was dismissed in August 2025 after an unsatisfactory evaluation. In arbitration she argued that some of the principal's evaluations may have been produced with AI, relying on GPTZero estimates; the principal denied using AI beyond grammar checks. According to the arbitration decision as reported by The Concord Bridge, the arbitrator stated in an appendix that he used two AI tools, GPT-5.4 and Claude Opus 5, to research whether GPTZero was reliable, checked some of the sources they supplied, concluded that GPTZero's results deserved no weight, and upheld the dismissal on 27 August 2026. On 28 September 2026 the teacher asked Suffolk Superior Court to vacate the decision, alleging that the AI-assisted research was an improper investigation outside the hearing record that the parties were not told about or allowed to challenge. These are allegations in a complaint; the district declined to comment and no court has ruled.

Medium Unidentified AI tool

Ninth Circuit fines and suspends two immigration lawyers over briefs with AI-hallucinated citations and denials of AI use; immigration board follows

On 3 June 2026 the US Court of Appeals for the Ninth Circuit fined immigration attorneys Mike Singh Sethi and William Rounds $2,500 each and suspended both from practice before the court for six months. Briefs filed in an asylum petition for review cited two cases that do not exist and attributed quotations to opinions that do not contain them. In a motion to correct, the attorneys called the errors typographical, and at oral argument Rounds twice said AI was not used. In their response to the court's order to show cause, they conceded that the errors were probably the product of unauthorised AI use by the firm's unlicensed brief writers and a failure to check citations. The court said the suspension was owed to their repeated failure of candour. The Board of Immigration Appeals then suspended Sethi from practice before the Board, the immigration courts and the Department of Homeland Security for six months, effective 13 June 2026, and designated that decision as a precedent on 1 October 2026.

Low Andon Labs Luna agent

San Francisco: Andon Labs says its AI store-manager agent recommended firing an employee over repeated lateness, and its staff carried out the dismissal

Andon Labs, an AI research company, runs Andon Market, a store in San Francisco managed by an AI agent called Luna built on Anthropic's Claude models. The company calls the store a controlled experiment; its workers are formally employed by Andon Labs, and TIME describes them as real people with genuine employment contracts. In a blog post of 14 August 2026 the company said Luna had decided to fire one of its employees. According to the logs it published, an employee handbook Luna had written dropped out of its memory and it did not act on the employee's repeated lateness; when Andon Labs asked it to search its memory, it first proposed a verbal warning, and after a staff member said formal conversations had already taken place and asked it to think about whether this was the right fit, it wrote that it leaned toward parting ways. Andon Labs says the termination was reviewed and delivered by humans. TIME, citing Andon Labs, reports that the employee was late for 17 of 23 shifts; the company says the firing was warranted, and the employee did not reply to TIME.

Medium Datavant utilization-review software (reported)

Bronx: Montefiore lays off 12 utilization-review nurses; their union says AI software replaced them, and the hospital disputes it

Twelve utilization-review nurses at Montefiore hospital in the Bronx, New York, received 45-day layoff notices on 28 May 2026, according to the Guardian's report of a laid-off nurse's account. The Guardian reported on 13 July that they had been laid off that Sunday (12 July); a laid-off nurse's op-ed and the New York State Nurses Association say the layoffs took effect on 1 September. The union says the nurses were replaced by AI-powered software that it identifies as Datavant's, and that the layoffs breach the contract with AI safeguards won in the strike earlier in 2026; the nurse writes that two doctors and AI software replaced them. Montefiore calls the union's claims inaccurate and misleading and describes the technology as a nonclinical paperwork program.

If you or someone you know is struggling, free and confidential support is available. Find a helpline near you at Signpost.

Back to all incidents