META has halted a controversial employee-tracking program after a security review revealed that sensitive data collected, including keystrokes and screen content, was more accessible internally than intended. The program, part of Meta's Model Capability Initiative, aimed to train AI by logging employee activities, but raised substantial privacy and security concerns. Reports indicated that the collected data was accessible across multiple internal tables, leading to a backlash from employees.
Meta's response included scaling back the project amid worries about compliance with privacy regulations and the potential reputational damage from breaking employee trust. The incident highlights the risks of monitoring programs and the complexities of data management.