Meta announced Wednesday that employees can now pause the collection of their mouse movements, clicks, and keystrokes for up to 30 minutes [1].
The move comes after significant internal backlash from staff regarding privacy concerns and the perception of invasive surveillance. Because the company uses this data to train artificial intelligence, the tension highlights the friction between corporate AI development and worker privacy rights.
The monitoring system is known internally as the Model Capability Initiative (MCI) [3]. The tool tracks activity across Meta's global workforce, including those in physical office locations and remote settings [4, 5]. Under the new updates, workers can request exemptions or use a feature to temporarily stop the data collection [1, 2].
According to an internal memo, these new controls allow staff to pause the collection for "up to 30 minutes at a time" [2]. The adjustment follows complaints that the software caused battery drain on devices and created an atmosphere of constant oversight [2, 6].
Stephane Kasriel, vice president of Meta's AI model-building Superintelligence Labs unit, said the team behind the software had also introduced several optimizations to reduce its impact [1].
The scale-back represents a compromise for the company as it seeks to gather high-quality human interaction data for its AI models without alienating its workforce [2, 6]. The 30-minute window is the maximum duration employees can opt out of the tracking [2].
“Meta is scaling back its employee-tracking programme, adding a feature that lets workers pause the collection of mouse movements.”
This policy shift indicates that Meta is struggling to balance the data-hungry requirements of its 'Superintelligence' AI goals with the privacy expectations of its professional staff. By granting a limited 30-minute window of privacy, the company is attempting to mitigate employee burnout and resentment while maintaining the core of its data-harvesting infrastructure.
