Technology
Meta adjusts employee evaluation rules to decouple performance from AI usage metrics
Internal guidance drops token tracking and AI adoption scores from review criteria while staff test a new autonomous agent.
The short version
- Meta notified staff that performance evaluations will no longer track AI tool adoption, token consumption, or internal AI usage designations.
- The policy change follows staff concerns and an ongoing lawsuit alleging that AI usage metrics unfairly penalized employees on leave.
- Workers are currently testing an experimental autonomous AI agent called Hatch on corporate devices ahead of a potential public rollout.
Key facts
- Meta released internal performance review guidance explicitly stating that adoption dashboards and token counts will not be used to evaluate employee impact.[Wired]
- A July lawsuit filed by roughly two dozen laid-off employees alleges Meta's previous AI usage metrics violated antidiscrimination laws by penalizing workers on medical or family leave.[Wired]
- Meta spokesperson Tracy Clayton stated that employee evaluations are based on overall contributions and denied that labels like 'AI Native' were ever used for performance reviews.[Wired]
- Meta employees have been testing an agentic AI tool named Hatch that can browse the web and navigate applications on work devices.[Wired]
What remains uncertain
- Whether Meta will issue a public release for the Hatch AI agent remains unconfirmed, as the company declined to comment on the internal testing.[Wired]
- The long-term workforce impact of AI productivity tools at Meta remains uncertain, amid employee fears of future cuts and conflicting statements regarding further layoffs.[Wired]