Meta Isn't Scrapping Its Controversial Employee Tracking Programme — Workers Just Won More Control

In April, Meta implemented software, the Model Capability Initiative (MCI), designed to monitor mouse movements and keystrokes carried out by employees to help train its AI models. This move prompted concerns among employees, particularly around privacy.

After receiving that feedback, along with other complaints including excessive data consumption that affected battery life and spiked internet usage, Meta announced in an internal memo that it would introduce changes to address concerns raised by workers.

According to Stephane Kasriel, vice president in Meta's AI model-building Superintelligence Labs unit, new controls will be introduced to address employee complaints. These include allowing staff to pause data collection for up to 30 minutes at a time, as well as the ability to request exemptions, according to a report by Reuters.

Kasriel also said engineers had implemented several optimisations to the software to address complaints about battery drain.

'While we remain confident in the privacy protections we put in place at launch, which went through several layers of risk review, we have heard your concerns about personal data on work devices, battery life, and wanting more control over when data capture occurs,' Kasriel said in the memo.

Beyond Clicks and Keystrokes

The intent of the tracking software was to help AI agents replicate human digital behaviour, including common tasks such as form-filling, menu navigation, and other routine workflows. However, some employees believe the scope of the tracking extends further, raising concerns about workplace surveillance and the potential erosion of privacy.

Under the current setup, employees have limited choice, as the programme is mandatory for those using company-issued devices. There is no general opt-out provision, although exemptions may be granted under specific conditions.

Data Extraction in Disguise?

The concerns raised by employees extend beyond performance issues. In another Reuters report, one employee shared findings from a detailed analysis of MCI log files conducted using Anthropic's Claude.

According to that analysis, MCI was alleged to be capturing more data than employees initially expected, including records of code changes, computer sleep and wake cycles, websites visited, and clipboard activity.

When the claims were raised internally, Meta management disputed the accuracy of the findings, although no detailed technical explanation was provided.

Meta has since announced improvements to its tracking software. The changes outlined so far focus primarily on work-related controls, including exemptions that could reduce the volume of data collected for AI training.

The issue could place Meta under heightened scrutiny in the European Union, given its existing challenges related to data collection and deployment. While US workers have limited protections against employer surveillance, companies operating under the European Union General Data Protection Regulation face far stricter requirements.

Organisations must establish a legal basis for processing personal data, clearly disclose what information is collected and meet heightened standards when handling sensitive data.

'This situation is not limited to Meta employees. It relates to every employee in every sector where they could be replaced. Everybody cares about this if they understand what it is,' Johnny Ryan, director of the Irish Council for Civil Liberties' Enforce unit, said in another Reuters report.

For now, Meta employees remain bound by company policy. As reported previously, roughly 7,000 workers were reassigned to train AI models.