Meta announced a plan called the Model Capability Initiative (MCI) in April 2026 to collect employee mouse movements, clicks, and keystrokes to train AI models [1, 2, 3, 4, 5]. By May, the tracking software was installed primarily on US-based employees' computers [3, 4].
The plan sparked strong opposition among Meta staff. Employees described the tracking as dystopian and issued a petition garnering more than 1,500 signatures opposing the data collection [1, 3, 5]. One worker said, "Having our actions train AI models felt very dystopian" [1]. Some criticized Meta as an "employee data-extraction factory," citing privacy and trust concerns [1, 4, 5].
Responding to the backlash, Meta’s Vice President of Superintelligence Labs, Stephane Kasriel, authored an internal memo around June 2–3 announcing new controls to scale back the program [1, 2, 3, 5]. Employees can now pause data collection for up to 30 minutes at a time and request exemptions from the initiative. Kasriel noted, "While we remain confident in the privacy protections we put in place at launch, which went through several layers of risk review, we have heard your concerns about personal data on work devices, battery life and wanting more control over when capturing happens" [1].
The development team also improved the tool’s efficiency to reduce its impact on laptop battery life and internet data usage following complaints [1, 2, 3, 5]. Meta said the data collected is only used to train AI models and includes safeguards to exclude sensitive content [1, 4]. Meta confirmed the software does not record screen content but focuses on interaction data from over 200 applications and websites used by employees [4]. A Meta spokesperson said, "The software is installed only on US employees’ devices, mainly to study computer interaction patterns, not to record screen content" [4].
However, European regulators and digital rights groups raised concerns about the indirect collection of data involving EU employees, especially regarding GDPR compliance. European digital rights lawyer Kleanthi Sardeli warned, "Even indirect collection of European employee data may raise GDPR concerns because using work communications in AI models is outside original purposes" [4]. There is some dispute over the tool’s scope, with some sources noting its effect may extend beyond US staff through intercepted communications [1, 3, 4, 5].
Meta has declined to comment publicly on the tracking program or the recent internal changes [1, 3]. The new employee controls and scaled-back collection mark the latest step after intense internal resistance toward the initiative.
