Meta Drops AI-Usage Performance Reviews, Rolls Out New Hatch Agent to Employees

Meta told employees this week it will no longer use AI adoption dashboards or token consumption to evaluate performance, walking back the 'AI-driven impact' framing it introduced a year ago that produced 'AI Native', 'AI First' and 'AI Enabled' worker labels. The reversal follows a July lawsuit from ~two dozen workers alleging the metric penalized staff on medical or family leave. Simultaneously, Meta is pushing a new internal agentic tool called Hatch — early testers report token usage keeps climbing anyway.