Meta pauses controversial employee monitoring to train AI models after data breach

Meta has indefinitely paused a controversial staff data-collection programme after an internal security failure left sensitive employee information accessible company-wide.

The data was collected under the Model Capability Initiative (MCI),  designed to train artificial intelligence (AI) models by logging the keystrokes, mouse clicks, and screen content of its US employees. An internal security notice revealed that misconfigured access control lists exposed employee activity across 45,000 hive tables, including full prompts, transcriptions, private conversations, and performance data. Meta stated there is currently no indication that the data was improperly accessed, and sources indicate the technical issue has been resolved.

However, the breach follows significant internal resistance to the monitoring, including an internal petition signed by more than 1,600 employees warning of security and regulatory risks. While executives, including CEO Mark Zuckerberg, previously defended the surveillance as essential, the company had already begun offering limited exemptions following worker protests. CTO Andrew Bosworth acknowledged that implementation fell short of the standards outlined in its privacy review, promising to share findings from an ongoing internal investigation.

£ - This article requires a subscription. 

Training Announcement: Find out more about our range of independent accredited qualifications from BCS and IAPP. These include professional certifications for practitioners covering data protection and AI governance, information security and risk management, along with access to public information.