Meta layoff AI: Inside the Human‑Driven Decision Process

The court documents reveal the contentious “Meta layoff AI” dispute, with plaintiffs accusing the firm of algorithmic selection. Twenty‑six former staff (↓ 26) claim AI tools flagged those on medical leave or with disabilities. Meta’s HR director Linh Doan responded in a filing, insisting that “human business leaders” applied objective criteria – performance scores, job level, tenure, location and managerial span – without reference to leave status or disability.

Meta layoff AI: How Selections Are Determined

According to Doan, the process begins by isolating a “cohort” defined by role and seniority. Once a business rationale is set, leaders must draft selection metrics that are quantifiable and tied to that rationale. For example, if the goal is to retain top talent, the metric might target employees with lower historical performance ratings.

Visa‑Sponsored Employees Targeted

Four workers on Meta‑sponsored visas were among those terminated. The filing explains that three were let go for “Consistently Met Expectations” – the lowest rating that does not fall below expectations – while the fourth faced a similar assessment despite being on parental leave. Doan emphasizes that leave status was not a factor.

“Decision‑makers could not deviate from the pre‑approved criteria,” Doan told the court.

The lawsuit also cites internal memos suggesting that AI adoption scores, tracked via the new Checkpoint system, feed into performance reviews. While Meta denies AI dictated layoffs, the link between AI usage and ratings remains unresolved. Analysts warn that tying performance to AI engagement could inadvertently penalize workers who take leave or have disabilities. For broader context, see recent Reuters coverage of AI‑driven HR practices.

Dispatch from Elara Vance (Night-Shift Breaking News Lead).