Meta Pauses AI Employee Tracking Program Following Data Leak
Meta has halted its AI model training program, Model Capability Initiative (MCI), after a data leak made sensitive employee information accessible to the entire company. The program had already raised privacy concerns among staff.
Meta has paused its controversial Model Capability Initiative (MCI) program, which tracked employee computer activity for AI training. The program had already sparked employee concerns prior to a significant data leak.
MCI was designed to automatically capture keystrokes, mouse movements, and screen content from employees' company laptops in the U.S. Its implementation in April 2026 occurred during a period of low employee morale following substantial layoffs.
The situation escalated in June 2026 when it was discovered that Meta had inadvertently left over 45,000 internal data tables, containing private conversations and performance records, accessible to all employees. Over 1,600 employees had already signed a petition demanding the program's cancellation, citing the company's failure to clarify data usage.
This incident highlights a broader trend where AI-powered surveillance tools are being deployed faster than necessary cultural discussions about trust and privacy. While the data breach gained headlines, the damage to employee trust began with the program's initial, opaque rollout.