Meta suspends employee tracking after internal revolt

Meta has effectively suspended a program designed to record the digital habits of its own workforce. The initiative, known as the Model Capability Initiative, aimed to capture data such as keystrokes and mouse movements to help artificial intelligence models better mimic human computer use. However, the project triggered a significant internal backlash, leading to its temporary halt after a data leak and intense employee resistance.

The move highlights a growing tension between tech giants seeking proprietary data to train advanced AI systems and the employees whose daily work is being monitored. While the company initially framed the effort as a way to improve AI performance in mundane tasks, many staff members viewed it as an invasive surveillance measure. The program is currently on hold, with Meta stating it will shift to an opt-in model if it is ever revived.

The push to automate white-collar tasks

The program originated in Meta’s elite AI research unit, a small group of researchers who believed that existing models lacked the ability to handle basic user interactions. These interactions include simple actions like selecting items from dropdown menus or using keyboard shortcuts. By harvesting this data, the company hoped to give its AI tools the dexterity needed to automate complex white-collar work, such as coding and administrative accounting.

The announcement in April made it clear that participation was mandatory. Meta’s chief technology officer, Andrew Bosworth, defended the initiative by emphasizing strict privacy controls, arguing that the data would not be made public. Despite these assurances, many employees felt uncomfortable with the idea that their every click and keystroke was being logged for the benefit of the company’s AI race.

Employees mobilize against mandatory tracking

Resistance to the program was immediate and widespread. Employees used internal channels to express their displeasure, with comments on the announcement often criticizing the lack of an opt-out option. The backlash quickly evolved into a grassroots movement, featuring shared memes and flyers that mocked the initiative. One prominent comment, which received high approval, expressed deep discomfort with the surveillance aspect of the project.

More than 1,800 staff members signed a petition demanding that the internal tracking stop. Organizers characterized the system as an intrusive form of observation, comparing it to a fictional all-seeing eye. This sustained pressure forced management to reconsider its approach, signaling that even within a tech giant, employees can exert significant influence over corporate data practices.

A leak forces a strategic pause

The program was ultimately derailed by a security incident later in the summer. An internal data leak exposed vulnerabilities in how the collected information was being handled, contradicting earlier claims of tight control. Following this breach, Meta halted the tracking program entirely.

According to reporting by GN technics/ai (en-US), the company stated that the initiative would remain suspended for at least two months. The shift suggests a move toward an opt-in framework, aligning with the demands of the protesting workforce. This episode underscores the broader challenge companies face when trying to balance the pursuit of technological advantage with the privacy expectations of their own teams.