Meta has halted its employee mouse-tracking initiative due to concerns about data security.

      The Model Capability Initiative, which tracks mouse movements and keystrokes to help train Meta’s AI, has been temporarily suspended after sensitive employee information was found to be accessible to all company personnel.

      On Monday, June 22, 2026, Meta announced it would pause the internal tool that monitors employee digital activities while an investigation is conducted into how a significant amount of confidential staff data became visible to everyone at the company.

      The initiative, known as the Model Capability Initiative (MCI), was launched in April 2026. As previous reports on the software indicated, it records the mouse movements, clicks, and keystrokes of U.S.-based employees, along with occasional screenshots, and uses this data to train Meta's AI models.

      The objective is to help AI systems understand how human workers typically perform tasks. However, this has resulted in a considerable amount of personal information being left exposed.

      This suspension comes after revelations from documents reviewed by Reuters that indicated sensitive employee data was unintentionally made accessible to all Meta staff. The exposed information included private conversations, performance metrics, and transcriptions—data that is sensitive enough when held by human resources, let alone being available across the entire workforce.

      The irony in this situation is striking: a program designed to analyze detailed employee work behaviors failed to protect that very information from the employees being monitored. The exposure was not caused by an external breach but rather by an internal permissions issue, representing a misconfiguration that opened a private dataset to the entire organization, for which Meta is responsible since it managed the data collection.

      Meta did not challenge the facts regarding the incident. “We have designed this program with privacy safeguards, and while we currently have no evidence that any data was improperly accessed by Meta employees, we are pausing it while we conduct an investigation,” stated spokesperson Tracy Clayton.

      The company did not specify the duration of the suspension, which leaves the program on hold indefinitely without a date for resumption.

      MCI has faced internal controversy since its inception, not only regarding security concerns but also due to employee protests against being monitored by software designed to learn from them—an issue that has been exacerbated by its introduction amidst job cuts.

      To alleviate some of the discontent, Meta later introduced a pause feature allowing employees to disable tracking for 30 minutes at a time, a concession that highlighted the extent of the constant monitoring otherwise.

      Legal implications have also become a significant concern, as logging keystrokes and capturing screenshots of identifiable employees conflicts with Europe’s data protection regulations. The initiative has raised red flags for potential violations of GDPR, which places stringent requirements on processing personal data and views workplace consent as unreliable due to the power disparities between employers and employees.

      A data leak that resulted in sensitive records being widely accessible exemplifies the failures these regulations aim to prevent.

      For the time being, mouse movements are no longer being recorded, at least not by MCI. Meta has stated it will investigate but has not disclosed the expected duration of this process or whether the initiative will resume in its current form, in a redesigned format, or not at all. These details, much like the data that triggered the pause, remain confidential.