Meta Drops AI Usage Quotas for Staff, But New 'Hatch' Agent Drives Token Spike
Meta is ending its practice of evaluating employees based on their AI tool usage, according to internal announcements. The company now says performance reviews will focus on overall impact, not AI adoption metrics. Meanwhile, employees testing the new agentic AI tool Hatch are consuming tokens at high rates.
Meta’s shift away from AI-usage metrics follows months of internal friction, including a leaked leaderboard that ranked employees by token consumption and a lawsuit alleging that usage-based evaluations unfairly targeted workers on leave. The company had previously rationed AI access after usage spiked, and now frames its new agentic tool, Hatch, as an optional experiment rather than a performance requirement. Employees testing Hatch remain wary of linking it to personal accounts, citing privacy and error risks. The change restores traditional impact-based reviews, though token consumption still climbs as staff explore the tool.
This policy reversal could ease workplace anxiety around AI adoption, but it may also slow the company’s internal push to normalize agentic tools. If Hatch becomes public, its early corporate testing—and the privacy concerns raised—could shape how society views autonomous AI assistants. Broader adoption may depend on trust, not quotas, affecting both employee morale and user acceptance of AI-driven automation in daily tasks.