Meta Halts Employee Tracking Program After Data Breach
Meta has halted an employee monitoring initiative that gathered computer activity for AI training purposes following an internal data breach and increasing pushback from staff. Source reported that Meta’s Model Capability Initiative gathered data including employees’ keystrokes and mouse movements to assist in training AI models to comprehend how individuals interact with computers and perform routine tasks, such as utilising dropdown menus and keyboard shortcuts. Meta initiated the program within its AI division earlier this year, with an announcement made to employees in April.
“While AI models excel at research and technical skills like coding, they still lack some of the basic ways that humans use computers like choosing from dropdowns and keyboard shortcuts,” an internal announcement read. Workers were initially not provided with the opportunity to opt out, leading to a significant backlash. Subsequently, over 1,800 employees endorsed a petition advocating for the termination of the program. Employees expressed apprehensions that the tracking system might inadvertently collect sensitive information, such as passwords, immigration documents, and credit card details.
Concerns have been raised regarding the potential for this data to facilitate the creation of AI systems that might supplant human labour. In June, Meta implemented a policy permitting employees to temporarily suspend tracking for a duration of up to 30 minutes while engaged in sensitive tasks. Employees opposing the program argued that the measure fell short of expectations. “We demand, and will continue to demand, that Mark and his leadership team scrap MCI entirely, immediately, forever”, they wrote, referring to Meta CEO Mark Zuckerberg.
The program was subsequently halted after a Meta engineer discovered that data collected through MCI could be accessed across the company, including private conversations. Meta stated that the program had been suspended, rather than permanently cancelled, and suggested that participation would be optional if MCI were reinstated. Meta has not provided clarity on the utilisation of the data or whether it was ultimately employed in the training of AI models. Source reported that the program has been suspended for two months following its halt.
