When AI Helps Decide Who Gets Laid Off: A Warning for Employers
A new lawsuit against Meta offers an early warning for employers using artificial intelligence to evaluate employees or make workforce decisions: An employer cannot avoid responsibility for a discriminatory decision by attributing it to an algorithm.
Twenty-six Meta employees have filed a federal lawsuit alleging that the company used AI-assisted systems to identify workers for layoff based on factors such as productivity, performance ratings and AI usage. According to the lawsuit, those measures placed employees with disabilities—or employees who had taken medical, parental or family leave—at an inherent disadvantage because they had less opportunity to generate the activity being measured.
Meta denies the allegations and maintains that its layoff decisions were made by people, not artificial intelligence.
The case is still at an early stage, and the employees’ allegations have not been proven. A federal judge recently declined to stop their scheduled layoffs, finding insufficient evidence of irreparable harm at this point. The court reportedly concluded, however, that the employees had raised serious questions about the role AI may have played. That ruling did not resolve the underlying discrimination claims.
The broader lesson for employers is already clear: AI does not create a safe harbor from existing employment laws.
An employer may include an employee who is on protected leave in a legitimate, nondiscriminatory reduction in force. Protected leave is not an absolute guarantee of continued employment. But an employer generally cannot select someone for termination because that person took protected leave—or use criteria that effectively penalize the employee for taking it.
That concern becomes particularly significant when an automated system measures output, attendance, communications or other workplace activity without accounting for periods of protected absence or the effects of a disability. A seemingly neutral metric can become a proxy for legally protected circumstances.
Employers using AI in hiring, performance management or layoffs should understand the data being measured, test the results for unintended disparities, account properly for protected absences and workplace accommodations, and ensure that meaningful human review actually occurs.
Ultimately, the employer—not the algorithm—remains responsible for the employment decision.
The Meta litigation may become an early test of how established discrimination and leave laws apply to AI-assisted layoffs. It almost certainly will not be the last.