Meta pauses an AI training program that tracks employees' keystrokes after a companywide leak (www.businessinsider.com)

Meta has paused its internal AI training program, the Model Capability Initiative (MCI), following a significant data leak that exposed sensitive employee information, including private conversations and performance analytics. The incident, classified as SEV 2, raised alarms internally as employees expressed frustration over the lack of adequate data protection measures. Meta confirmed the ongoing investigation but noted that there is currently no evidence of malicious access to the leaked data, emphasizing that privacy safeguards were in place when the program was designed. This incident is particularly significant for the AI/ML community as it highlights the ethical and privacy concerns associated with using employee data for AI model training. Mandatory participation in MCI had already drawn criticism from employees worried about their keystrokes and mouse movements being recorded. This breach underscores the challenges organizations face in balancing the advancement of AI technologies with the protection of individual privacy, raising questions about data governance in AI initiatives. As Meta navigates this issue, it reflects broader concerns about security and trust in AI deployment across the tech industry.