Meta Pauses Controversial Employee Tracking Program After Internal Security Breach Exposes Sensitive Staff Data

Meta has halted a controversial employee tracking program after an internal security flaw allowed other company personnel to access potentially sensitive information collected through the initiative.

“We designed this program carefully with built-in privacy safeguards, and while we have no evidence at this point that any Meta employee accessed the collected data improperly, we are pausing the initiative as we run a full investigation,” said company spokesperson Tracy Clayton.

Meta first rolled out the Model Compatibility Initiative (MCI) tool to U.S.-based employees in April. According to workers who have organized against the tool over privacy, cybersecurity, and personal autonomy concerns, the system “collects granular computer inputs including mouse movements, click locations, and keystrokes, alongside full screen content.” When MCI launched, employees had no option to opt out of participation; the policy was only adjusted to allow limited opt-out after widespread staff protest.

| Got a Story Tip? |

|---|

| Are you a current or former Meta employee hoping to share insight into company operations? We want to hear from you. Using a personal non-work phone or computer, contact our reporter securely via Signal at @Peard33.24. |

Meta executives have repeatedly defended the data-gathering project, arguing it is a necessary step to train AI systems to interact with computer software the way humans do, and that employee activity provided the best example dataset for the artificial intelligence to learn from.

This Monday, a Meta engineer published an internal security notice confirming that databases storing all information gathered by MCI had been left open and accessible to any person inside the company.

A former employee actively involved in organizing pushback against MCI described the security lapse as “a total mess” — and one that workers had predicted would happen. “When staff raised concerns, leadership doubled down and refused to acknowledge the risks workers flagged for the privacy and security of both employee and customer data,” the person said. “Leadership has clearly built an authoritarian work environment where employees are no longer respected or listened to.”

After frustrated, critical comments about the security issue flooded internal company forums on Monday, Meta’s decision to fully pause MCI caught many staff off guard. The company shared news of the pause with WIRED several hours before announcing the change to its own employees. In the interim, multiple workers told WIRED they were confused because the tool was still actively running on their work laptops.

Late Monday, Stephane Kasriel, Meta’s vice president overseeing AI research, formally announced the pause to staff. He confirmed the security issue was first discovered on June 18, and the initial exposure was addressed within four hours. However, the first patch did not hold, and teams had to implement additional restrictions to further lock down data access. The flaw made “some MCI-derived data” accessible to more people than intended, he wrote, without sharing further details on the scope of the exposure.

“We will only re-enable MCI once we are confident in the effectiveness of our data protection controls,” Kasriel said. He added that Meta will share more updates on MCI’s future soon, as the program has already “gathered sufficient data to assess the long-term value of the tool.”

Zoë Schiffer contributed to this report.

---

Update 12:15am EDT, 6/23/2026: This story was updated to include new details from Stephane Kasriel’s internal note to staff and additional comments from employees.