Meta Halts Worker Tracking AI Program After Staff
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Meta paused its Model Capability Initiative (MCI) on Monday after realizing collected employee data was potentially accessible to anyone inside the company, though a spokesman said there is "no indication at this time that any data was improperly accessed by Meta employees."
- Nearly 2,000 Meta workers signed a petition demanding the MCI program be cancelled, and employees openly insulted management using explicit language in an internal meeting over AI-driven organizational changes, per Wired.
- Meta's initial compromise — letting workers opt out of tracking for up to 30 minutes at a time — was dismissed by one current employee as "just an attempt at damage control."
- Meta is spending up to $145bn (£109bn) this year on AI and has run extensive layoffs and reorganizations around AI initiatives, fueling the employee anger behind the MCI backlash.
- Current employees told the BBC they are open to improving AI models and competing with Anthropic and OpenAI, but objected that tracking "was forced on us, there was no consent," with one saying "I've never seen morale here so bad."
- A former Meta employee described the company's AI catch-up push as "chasing your tail," calling the direction "exhausting and depressing."
Why it matters: The pause exposes a tension inside Meta's $145bn AI push: the company wants to compete with Anthropic and OpenAI but is alienating the very engineers who build its models — one employee said morale is the worst they've seen, and nearly 2,000 workers signed a petition. The data-access gap also raises concrete questions about how Meta will collect the sensitive behavioral signals needed for frontier AI training without consent.
