What happened when Meta tried to collect employee data
Executive Take
Mandatory employee surveillance for AI training can trigger faster, more organized backlash than external product controversies. Leaders should build opt-in consent and data access controls before launch, not after a leak forces a retreat.
Executive Summary
Meta launched the Model Capability Initiative, a mandatory program collecting employees' keystrokes and mouse movements to train AI models. Over 1,800 employees signed a petition opposing it. After a leak revealed the data was accessible company-wide, Meta suspended the program within two months.
Why It Matters
HR and technology leaders should care because this shows white-collar employees will organize quickly against AI-driven monitoring, even using labor law protections. It also shows internal data governance failures can undo an AI initiative faster than any backlash.
Bizquad Perspective
The real failure wasn't the surveillance idea but Meta's own data controls, which exposed private conversations company-wide and handed critics the exact proof they needed.