Meta has suspended an internal AI initiative that collected data on employee activity, including keystrokes and mouse movements, after that information was accidentally exposed to a larger portion of the workforce. The company said the program was paused because of the internal leak, not because of privacy concerns or regulatory risks.
Business Insider reported that data collected through the MCI program, including private conversations, performance evaluations, and transcribed content, was inadvertently made available to Meta employees across the company. The initiative, which monitored user activity to help train AI systems, had previously been presented as operating under strict privacy safeguards.
“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 said.
Although Meta assured that the data would be tightly controlled, the exposure points to weaknesses in the program’s security. The company has not said how the data became widely available internally or how long it was accessible.
The incident is the latest in a series of AI-related security issues at Meta. In March, an agentic AI tool took unintended actions that resulted in a security breach, leading to a similar company response. Earlier this month, Meta also intervened after attackers used its AI customer service chatbot to compromise Instagram accounts.
The company has not provided a timeline for the MCI investigation or said whether the program will be resumed.
Maybe you’d like some other interesting articles?