Meta pauses worker tracking program after internal security breach concerns

Meta has paused its internal Model Capability Initiative after weeks of employee backlash and a security problem that exposed sensitive information to employees across the company. The program, which began in May 2026 on U.S.-based employees’ computers, tracked mouse movements, clicks and keystrokes as part of Meta’s push to build AI agents that can carry out work tasks autonomously.

The tracking system sat at the center of a widening conflict inside the company over informed consent and workplace surveillance. In an internal memo, Stephane Kasriel, a vice president in Meta’s Superintelligence Labs unit, told workers they would be able to pause data collection for up to 30 minutes at a time and request exemptions. He also said the software team had made battery-life optimizations after employees complained that the program was using so much data it was increasing home internet usage.

Meta said the launch had gone through several layers of risk review and that it remained confident in its privacy protections. A Meta spokesperson declined to comment.

Inside the company, the backlash turned caustic. Workers nicknamed the initiative an “Employee Data Extraction Factory,” a label that captured the growing unease around how much visibility the company was seeking into everyday computer use while building its AI products.

The pause followed a separate internal leak reported on June 22, 2026, after a worker flagged a security issue. According to the reports, sensitive material including private conversations, performance data and transcriptions had been inadvertently made available to the entire Meta staff. That breach transformed an internal policy dispute into a broader accountability problem, raising questions about who had access to employee data, how it was governed and whether the controls were adequate for a company building products around trust and privacy.

The episode also lands at a sensitive moment for Meta’s larger AI restructuring. The company is racing to expand its AI ambitions, but the pause suggests there are limits to how aggressively it can repurpose employee activity as training material without triggering internal resistance. It also threatens to intensify regulatory scrutiny in the European Union, where Meta is already facing legal clashes over how tech companies collect and use data.

For Meta, the immediate question is not only whether the tracking program returns in a revised form, but whether the company can persuade workers, regulators and the public that internal monitoring in the name of AI development will not erode the privacy boundaries it claims to protect.