Meta Employee Surveillance
- Reports say Meta installed software on employee machines to record keystrokes, mouse movements and site usage for an internal AI project. - The tracking allegedly covered sites like Google, LinkedIn and Wikipedia, and employees have protested the monitoring. - Critics argue the telemetry is being converted into training signals that could teach agents to replicate human tasks, sparking internal backlash (cnbc.com).
Meta is installing software on U.S. employees’ work computers to record clicks, keystrokes, mouse movements and some screen content for artificial intelligence training. (reuters.com) The tool is called Model Capability Initiative, or MCI, and Reuters reported on April 21 that it will run on work-related apps and websites. Meta spokesperson Andy Stone said the data will be used for model training, not performance reviews. (reuters.com) CNBC reported on April 22 that Meta plans to monitor activity across hundreds of sites and apps, including Google, LinkedIn, Wikipedia, GitHub, Slack and Atlassian products. CNBC also said the internal list originally included OpenAI’s ChatGPT and Anthropic’s Claude before the list changed. (cnbc.com) The basic idea is simple: watch how a person uses a computer, then turn those actions into examples for an AI system. Meta said it needs records of things like selecting from dropdown menus, clicking buttons and using keyboard shortcuts so its agents can learn those tasks. (reuters.com) That project sits inside a broader company push to use AI agents for office and coding work. In a separate memo, Chief Technology Officer Andrew Bosworth told employees Meta’s “AI for Work” effort had been rebranded as Agent Transformation Accelerator, or ATA. (reuters.com) Bosworth wrote that Meta’s goal is a workplace where agents “primarily do the work” and employees direct, review and improve them. Reuters reported that he also said the company would be “rigorous” about building data and evaluations from daily work interactions. (reuters.com) Employees pushed back internally after the memos circulated. CNBC reported that a Meta Superintelligence Labs memo was sent to calm privacy concerns, and discussion of the tracking spread on internal chat boards. (cnbc.com) Meta says the collection has limits. Stone said safeguards are in place to protect “sensitive content,” but Reuters reported that Meta did not explain publicly which categories of information will be excluded. (reuters.com) The timing is tied to Meta’s larger race to catch up in generative AI. CNBC reported that Mark Zuckerberg has spent heavily since last summer, brought in Scale AI co-founder Alexandr Wang to lead Meta Superintelligence Labs, and unveiled the Muse Spark model on April 8. (cnbc.com; cnbc.com) So the dispute inside Meta is not just about monitoring software. It is about whether ordinary employee work on a company laptop becomes raw material for systems Meta hopes can later perform more of that work on their own. (reuters.com; cnbc.com)
