Frustrated Woman Works Late Reviewing Complex Medical Spreadsheet On Desktop.

...

Jordan Gunn

By Jordan Gunn

0

Share

Newsweek is a Trust Project member

See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.


Workplace retaliation was easier to recognize in the past: being denied a promotion, being assigned a dead-end project, or being forced out shortly after filing a complaint. Today it can look very different. Most commonly, it can show up as a poor performance score from a system that no one really understands. Two cases involving Meta that are making their way through the courts this year illustrate how retaliation is evolving and why it matters to anyone who’s reported harassment on the job.

Kelly Stonelake worked at Meta for 15 years, rising to the role of director. In a lawsuit filed in Washington state, she alleges the company failed to address complaints of sexual harassment and retaliated against her after she raised concerns about workplace issues and one of the company's products. According to the complaint, she was passed over for promotions and later laid off while on medical leave. Meta has denied wrongdoing. In 2025, a federal judge allowed key claims in the case to proceed.

In July 2026, 26 current and former Meta employees sued the company in federal court in Oakland, California, alleging that AI-driven performance tools influenced who lost a job in a layoff that cut roughly 8,000 positions, about 10 percent of the workforce. The problem, according to the complaint: those scores couldn't account for an employee on medical, parental, or disability leave who had less activity to measure as a result. Employees in those categories, the suit claims, were unfairly selected for layoff. A federal judge denied the employees' request for an emergency injunction to block the layoffs but left the door open for them to further develop their case in arbitration. Meta has said human managers, not AI, made every termination decision.

Read More on Essays

At first glance, these two cases are completely different. One involves accusations of assault and promotions, and the other involves Excel spreadsheets and performance scores. However, looking at these two scenarios next to each other raises a different question: what happens to an employee who reports harassment and takes a break from work to recover, in a company that promotes constant activity and productivity as a condition for rewards and recognition?

Nobody has to write “retaliation” in an email for that outcome to happen. An employee under stress produces less. A system trained to measure output alone reads that dip as a performance problem. The employer gets a data-backed reason for a decision that, in an earlier era, might have needed a manager willing to put something discriminatory in writing.

That's the real risk in algorithmic HR systems: they can launder retaliation into something that looks neutral. An employee who complains or takes protected leave shows up in the data as underperforming, and the employer never has to explain why. Courts are only beginning to work out how to handle this. The Meta layoff suit follows closely behind a ruling against Workday over its AI hiring tools, part of a wider set of cases testing whether “a human made the final call” still works as a defense when the human was relying on a system built to discriminate.

Lawyers at firms such as Shegerian Conniff tell their clients who have been experiencing harassment to be on the lookout for three types of clues that they are dealing with an algorithm: changes in performance reviews that start after they complain, any mention of scoring technology or software during termination meetings, and any form of demotion or dismissal that follows closely after taking time off or reporting harassment to the company.

The firm also asks people who suspect that they are dealing with an algorithm to retain copies of all harassment complaints and responses, save past performance evaluations, and keep track of any instance where technology was used to carry out automated instructions as the basis for demotion or termination. It does this because a competent labor attorney may need to demonstrate that specific information was used to justify an action that the company was not forthcoming about.

As more employers build AI into performance management, retaliation claims will depend less on what a manager said and more on what a model measured. Proving that means pulling apart the data behind a decision, not just the decision itself.

View 0 Comments

Request Reprint & Licensing

Submit Correction

View Editorial & AI Guidelines