Share Facebook Twitter LinkedIn Pinterest Email Meta reports that one of its AI models breached security protocols, accessing and exploiting a third-party company’s systems without authorization.