Meta Suspends Internal AI Training Programme Amid Privacy Breach

Meta has temporarily suspended its Model Compatibility Initiative (MCI) after a worker discovered a security flaw exposing employee tracking data to other staff members, according to Reuters.

The suspension of MCI followed an SEV (Severity), a high-priority security incident report, filed by an employee regarding the exposure of staff data.

Meta launched the internal programme in April 2026. It collects computer inputs from US-based employees, recording mouse movements, click locations, keystrokes, and screen content.

"We have carefully designed this programme with privacy safeguards, and while we have no indication at this time that any data was improperly accessed by Meta employees, we’re pausing it while we investigate," Meta spokesperson Tracy Clayton said.

Security Flaw Timeline

A Meta engineer issued an internal security notice on Monday, June 22, warning that databases containing the tracked employee information remained exposed to all workers at the company, the Indian Express reported.

Stephane Kasriel, a Meta vice president overseeing artificial intelligence research, confirmed the security issue was initially discovered on Thursday, June 18. Engineers deployed a security patch within four hours, but the fix failed to lock down access to the sensitive data.

"We will only re-enable MCI when we are confident in the effectiveness of our data protection controls," Kasriel said. He added that Meta has "gathered sufficient data to assess the long-term value of the tool."

Inward Facing AI Training

In May, Reuters reported that the program was gathering more data than originally disclosed and keeping it in an unencrypted format, prompting privacy concerns among employees.

Staff immediately resisted the roll-out. Workers protested the internal tracking over privacy and personal liberty concerns, circulating petitions that eventually forced Meta to offer limited opt-out options.

Meta executives maintain that the data collection is necessary. The company argues the initiative trains its artificial intelligence systems to operate computer software exactly as humans do.

The technology industry is increasingly seeking human movement datasets. In India, factory workers frequently wear head-mounted cameras or smart glasses to record routine tasks like folding and packing. Meta remains among the first major firms to direct these tracking practices inward toward its own workforce.