Meta Faces Internal Revolt Over AI Surveillance Tool

Meta employees are pushing back against the company's Model Capability Initiative, a tracking tool designed to monitor mouse movements and keystrokes for AI training. The program has drawn sharp criticism from staff who label the firm an employee data extraction factory, while raising potential conflicts with European privacy laws.

The initiative, which captures granular activity including clicks, dropdown navigation, and clipboard contents, has sparked significant technical and ethical friction. Workers report that the software consumes massive amounts of bandwidth, exhausting home internet quotas in mere days. Beyond performance issues, internal concerns persist regarding the scope of surveillance, which reportedly extends to code modifications, device sleep cycles, and the contents of direct communications.

Meta spokesperson Dave Arnold defended the program, dismissing internal criticisms as inaccurate. He maintained that the tool is restricted to U.S.-based hardware and that non-U.S. staff were notified only because they interact with American colleagues. However, legal experts suggest the cross-border collection of email and chat data could trigger violations of the European Union’s General Data Protection Regulation. This tension unfolds as the company navigates a broader industry shift toward AI-driven restructuring, following a year marked by significant headcount reductions across the tech sector.