Meta is facing a lawsuit alleging that its AI-based layoffs process disadvantaged employees who had taken legally protected leave, raising fresh questions about the reliability of automated workforce decisions.
Filed in July 2026 by 26 current and former employees, the suit claims Meta used what it calls "a constellation of internal artificial-intelligence systems" to score workers ahead of a reduction affecting roughly 10% of its workforce in May 2026. The scoring inputs reportedly included performance reviews, calibration scores, productivity and output metrics, so-called "AI-native" assessments, and even AI token usage.
The plaintiffs argue these metrics inherently accumulate with time actively worked — meaning employees on protected leave (parental, medical, disability) could never generate comparable data and were structurally scored lower. Specific cases cited include a researcher flagged for layoff while on parental leave, a manager demoted after medical leave and then targeted during a second medical leave, and an engineer whose performance rating was lowered because an injury interrupted continuous work.
The suit alleges violations of the Americans with Disabilities Act, the Family and Medical Leave Act, the Pregnant Workers Fairness Act, and Title VII — arguing Meta failed to "neutralize" protected-leave status as an input to its AI systems. Meta has denied the claims, stating that "workforce and organizational decisions were made by people, not AI, and that continues to be true."
Beyond the labor dispute, this case is a warning about any system that scores people or campaigns automatically using engagement, output, or productivity data. If the underlying data structurally disadvantages a group — such as employees on leave, or audiences inactive during a specific season — the algorithm will amplify that bias rather than correct it. Any organization using AI-based scoring or targeting tools should build in data-gap correction and human review before those outputs drive high-stakes decisions.
The case is also a reminder that "the AI decided" is no longer a shield from accountability — it can be a liability. Even Meta emphasizes that a human made the final call. Brands and companies deploying automated decision systems should document how conclusions are reached and keep a human-in-the-loop, both to manage regulatory risk and protect reputation.
For guidance on responsible AI governance in your marketing operations, explore Best Partner's services or get in touch.
Twenty-six current and former Meta employees allege the company used a set of internal AI systems to score workers ahead of a workforce reduction, and that the scoring method disadvantaged employees on protected leave such as parental or medical leave.
The complaint cites the Americans with Disabilities Act, the Family and Medical Leave Act, the Pregnant Workers Fairness Act, and Title VII, arguing Meta's AI scoring failed to account for employees who could not accumulate performance data while on legally protected leave.
Any AI system that scores people or campaigns using activity or output data can encode structural bias against those with legitimate data gaps. Building in human review and correcting for known data gaps before acting on AI outputs reduces both fairness and reputational risk.
To apply what you just read to your own site, start with a free audit of where things are now.
A strategist replies within 24 hours on business days.
Opening an AI before a search engine is becoming the default. As the path back to original creators narrows, where does competitive advantage in the AI era actually come from?
Meta expanded its vibe coding app Pocket to all U.S. users. Describe what you want and it builds a small app or game — then that output circulates in a social feed.
Meta's CEO published a 6,500-word essay on personal superintelligence that doubles as an argument against AI regulation — while Meta cut 20% of staff earlier this year.