Meta Sued for Allegedly Using Discriminatory AI in Layoff Decisions (gizmodo.com)

Twenty-six former employees of Meta are suing the company, alleging that its use of AI-powered systems in layoffs was inherently discriminatory. During a significant layoff in May that affected 8,000 employees, the lawsuit claims that Meta used internal AI tools, including a large-language model called Metamate and algorithmic scoring based on productivity metrics, to determine which employees would be terminated. This method allegedly led to the disproportionate selection of employees on medical leave, including those who were pregnant or had recently taken family leave, as the AI systems did not account for protected absences and instead penalized them based on performance metrics. The implications of this lawsuit are significant for the AI and machine learning community, raising concerns about the ethical use of AI in decision-making processes, particularly in high-stakes scenarios such as layoffs. The plaintiffs argue that these AI systems, which relied on potentially biased data inputs—including keystrokes and work-related communications—failed to provide a fair evaluation of employee performance. Meta has denied the allegations, asserting that human judgment, rather than AI, was central to workforce management decisions. This case not only challenges the accountability of AI systems in corporate environments but also highlights the urgent need for safeguards against discrimination in AI methodologies.