Meta has suspended an internal programme designed to collect employee computer activity data for artificial intelligence training after a security issue exposed information to staff across the company.
The programme, known as the Model Compatibility Initiative (MCI), was introduced in April to gather data such as mouse movements, clicks, keystrokes and screen interactions to help train AI systems capable of operating software in a more human-like manner.
According to reports, the initiative was initially deployed among employees in the United States and faced criticism from staff over privacy and data protection concerns.
Meta confirmed that it has paused the programme while investigating the security issue.
“We have carefully designed this programme 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,” a company spokesperson said.
The suspension followed an internal security notice indicating that databases containing information collected through the programme were accessible to employees who were not authorised to view the data.
Meta executives said the vulnerability was first identified on June 18 and that engineers issued a security patch within hours. However, the company later determined that additional measures were required to fully secure access to the information.
Stephane Kasriel, a Meta vice president overseeing AI research, said the programme would remain suspended until the company is satisfied that adequate safeguards are in place.
“We will only re-enable MCI when we are confident in the effectiveness of our data protection controls,” he said.
The initiative forms part of a broader effort by technology companies to gather real-world behavioural data to improve AI systems and robotics. Meta has argued that the programme is necessary to train AI models to interact more effectively with software applications.
The company said it will provide further updates after completing its review and assessing the long-term value of the project.
