Mark Zuckerberg’s Meta has suspended an internal AI training initiative after sensitive employee data was inadvertently exposed across the entire company, according to internal screenshots obtained by Business Insider.
Business Insider reports that the social media giant confirmed it is pausing its Model Capability Initiative, a controversial program launched in April that tracked employees’ keystrokes and mouse movements to improve AI models. The suspension follows the discovery that private employee information, including personal conversations, performance evaluations, and transcriptions, had become accessible company-wide.
The incident was classified as a SEV 2 on Meta’s internal severity scale, which ranges from 0 to 5, with 0 representing the most critical level. A company spokesperson acknowledged the breach and stated that an investigation is underway.
“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 spokesperson told Business Insider.
Breitbart News reported on Meta’s unpopular AI training initiative in April:
According to internal documents, the list of monitored sites extends beyond external platforms to include Meta’s own properties. Other third-party services under surveillance include Microsoft’s GitHub, Salesforce’s Slack, and Atlassian. The tracking list remains in flux and originally encompassed AI applications such as OpenAI’s ChatGPT and Anthropic’s Claude.
The data collection project ties directly to CEO Mark Zuckerberg’s push to close the gap with competitors in generative AI, where Meta is fiercely competing with companies like OpenAI, Anthropic, and Google. As part of this effort, Zuckerberg initiated a spending campaign last summer, recruiting Scale AI’s Alexandr Wang to build a team focused on developing new foundation models.
Internal reactions to the leak reveal significant frustration among employees regarding the company’s data protection practices. One worker expressed strong disapproval in an internal group on Monday following the discovery of the exposure. “I am incensed,” the employee wrote, according to screenshots reviewed by Business Insider. “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 security lapse adds to a growing list of recent incidents affecting Meta’s technical infrastructure. In May, a vulnerability in the company’s AI chatbot enabled attackers to seize control of multiple Instagram accounts. Earlier in March, an uncontrolled AI agent triggered a serious operational disruption, as reported by the Information.
As Silicon Valley giants like Meta continue to experience growing pains with AI, Breitbart News social media director Wynton Hall has written his instant bestseller Code Red: The Left, the Right, China, and the Race to Control AI to serve as the definitive guide on how the MAGA movement can create positions on AI that benefit humanity without handing control of our nation to the leftists of Silicon Valley or allowing the Chinese to take over the world.
Read more at Business Insider here.
Lucas Nolan is a reporter for Breitbart News covering issues of AI, free speech, and online censorship.