Meta has d its Model Compatibility Initiative that tracked employee mouse movements, clicks, keystrokes, and screen content to train AI agents, after some of its collected data became accessible to more employees than intended. Meta says it has no evidence the information was improperly accessed and will not restart the program until it is confident in its safeguards. Wired reports: Meta rolled out the Model Compatibility Initiative (MCI) tool in April to US employees. The tool "collects computer inputs such as mouse movements, click locations and keystrokes, as well as screen content," according to workers who have been petitioning against it over privacy, security, and personal liberty concerns. When MCI launched, employees couldn't opt out, but that changed to a limited degree after workers protested. Meta executives have repeatedly defended the data-gathering project, saying it was necessary to train AI systems to operate computer software the way humans do and that employees were the best examples for the artificial intelligence to learn from. On Monday, a Meta engineer issued an internal security notice stating that databases filled with information gathered by MCI had been exposed to anyone inside the company. A former employee actively involved in pushing back against MCI describes the lapse as "a mess" -- and one that employees had expected would occur. "When workers raised concerns, leadership doubled down and failed to acknowledge the risks workers raised about the safety and privacy of worker and customer data," the person says. "Leadership has clearly created an authoritarian environment where workers are no longer respected or heard." But after critical comments poured into internal forums on Monday expressing frustration about the security issue, Meta shocked some of its staff by pausing MCI altogether, telling WIRED about the development several hours before announcing it to employees. A few workers told WIRED they were confused in the meantime because the tool was continuing to run on their laptops. Late on Monday, Stephane Kasriel, a Meta vice president overseeing AI research, announced the and told staff that the security issue had been discovered on June 18 and addressed within four hours. But the initial fix didn't stick and access to the data had to be further locked down. The issue made "some MCI-derived data" accessible to more people than intended, he wrote, without elaborating.Read more of this story at Slashdot.
Meta’s Program That Spies on Every Employee’s Computer Just Blew Up in Its Face in Spectacular Fashion