Meta retreats on employee AI tracking amid internal revolt
New controls allow staff to pause data collection for 30 minutes, marking a significant concession after backlash over privacy and resource usage.
Meta has revised its controversial plan to monitor employee computer activity for artificial intelligence training, introducing new controls that allow staff to pause data collection. The move follows significant internal criticism of the Model Capability Initiative (MCI), a tool designed to log keystrokes and mouse clicks to help train AI agents.
An internal memo from Stephane Kasriel, vice president of the Superintelligence Labs unit, confirmed the changes. According to the document, employees can now pause the data collection for up to 30 minutes at a time or request exemptions from the initiative altogether. Kasriel stated that the team had introduced several optimizations to reduce the tool's impact on laptop battery life, addressing concerns that the software was consuming excessive internet data for remote workers.
The reversal comes after weeks of backlash from Meta staff. When the company announced the MCI tool in April, it argued that real examples of computer usage were necessary to build helpful agents. However, employees described the tracking as dystopian, with fears that the surveillance was being imposed just as the company prepared for major workforce reductions. A petition against the tracking initiative has gathered more than 1,500 signatures.
Tensions at the company are heightened by ongoing restructuring. Meta has laid off approximately 2,000 employees this year and previously announced plans to cut 10% of its workforce, equating to roughly 8,000 staff. Critics within the organisation have argued that the tracking tool represents another instance of the company forcing AI adoption on staff despite their reservations.
While Meta maintained that the data was not used for any other purpose and that safeguards were in place to protect sensitive content, the company acknowledged the concerns regarding personal data on work devices. The revised policy marks a strategic pivot to quell internal dissent while continuing to gather the data required for its AI development pipeline.
