Meta pauses employee computer-activity tracking program used for AI training after internal data-access concerns
The facts
- Meta said it paused its internal Model Capability Initiative while it investigates data security concerns.
- The program was designed to collect employee computer-activity data to help train Meta's AI systems.
- Reports said the program tracked employee inputs including mouse movements, clicks and keystrokes.
- Meta halted the program after learning that sensitive employee data collected through it was accessible to workers across the company.
- The initiative was launched in April and applied to U.S.-based employees.
- The program prompted internal employee opposition centered on privacy and security concerns.
- More than 1,600 employees signed a petition opposing the data-collection program.
- Meta has not said how long the program will remain paused.
Context
What was the Model Capability Initiative?
It was an internal Meta program intended to gather data on how employees use computers so the company could train AI systems to perform software tasks more like humans do TechSpot,Independent.
What kinds of employee data were being collected?
Multiple reports said the program logged inputs such as keystrokes, mouse movements and clicks, and some accounts said it also captured screen content from work devices BBC,Indian Express,Independent.
What remains unresolved?
Meta says it is investigating the data-security issue and has said it has no indication so far that the data was improperly accessed by employees, but it has not said how long the pause will last Reuters,PC Magazine.
Common ground, and where they split
- Common ground
- Sensitive employee data was exposed too broadly inside the company, making Meta’s AI-training program a real privacy and security failure, not just an internal process dispute.
- Where Left and Right split
- Workers being made to bear privacy risks for an AI project, or a company enforcing proper limits on its own innovation efforts.
How left and right read it
Meta’s pause matters because the program was built to collect employee computer-activity data for AI training, and it was halted only after sensitive employee data was found to be accessible across the company. The internal opposition, including a petition signed by more than 1,600 employees, highlights a basic power imbalance: workers were being asked to absorb privacy and security risks in service of an AI project.
The key issue is institutional restraint: Meta paused the Model Capability Initiative once it was investigating data security concerns and learned sensitive employee data was accessible across the company. For a company building AI systems from employee computer-activity data, that matters because even internal innovation efforts need clear limits and trustworthy handling of sensitive information.
The receipts — 58 sources
Wire services (2)
Other coverage (56)
A news briefing built to show both sides.
The day’s biggest stories in one short brief, built to show both sides and the competing values behind the headlines. Free in your inbox.