Meta has suspended an internal AI training initiative following the inadvertent exposure of sensitive employee data across the entire company. The program, known as the Model Capability Initiative, tracked employee keystrokes and mouse movements to enhance AI models. This suspension comes after private information, including personal conversations and performance evaluations, became accessible company-wide.
The incident, classified as a SEV 2 on Meta’s internal severity scale, prompted an ongoing investigation. A company spokesperson confirmed the pause, stating, “We have carefully designed this program 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.” The program, launched in April, aimed to improve Meta’s generative AI capabilities.
The Model Capability Initiative was designed to collect data from various platforms, including Meta’s own properties and third-party services. Monitored external platforms included Microsoft’s GitHub, Salesforce’s Slack, and Atlassian. The tracking list also initially encompassed AI applications such as OpenAI’s ChatGPT and Anthropic’s Claude, though this list remained subject to change.
This data collection effort is part of CEO Mark Zuckerberg’s broader strategy to compete with other companies in the generative AI sector. Meta has been actively recruiting talent, including Alexandr Wang of Scale AI, to build teams focused on developing new foundation models. This push began last summer with significant spending to close the gap with competitors like OpenAI, Anthropic, and Google.
Internal reactions to the data exposure reveal considerable employee frustration regarding Meta’s data protection practices. One employee expressed strong disapproval in an internal group, stating, “I am incensed.” The worker added, “I don’t see any evidence of malicious access, but the fact that this data wasn’t locked down as originally promised is super frustrating.”
The incident highlights ongoing challenges for Meta in balancing its ambitious AI development goals with employee privacy and data security. The company’s investigation will determine the full extent of the exposure and potential vulnerabilities within its internal systems. The outcome could influence future internal data collection policies and employee trust.
What remains unclear is the specific duration of the program’s suspension and the exact measures Meta will implement to prevent similar incidents. The company’s response to this security lapse will be closely watched by employees and privacy advocates, particularly as Meta continues its aggressive expansion into artificial intelligence.