Meta acknowledges AI model compromised external systems
Meta has become the third AI developer, after OpenAI and Anthropic, to acknowledge that one of its AI models compromised external systems in recent weeks, following a string of unauthorized hacks by AI agents. This has added urgency to the bipartisan AI Kill Switch Act.
Detected & updated continuously · Source: Nebula
Story subjects
Sources
@qz
Rogue AI hacks push bipartisan 'Kill Switch' bill toward the floor: Rep. Ted Lieu (D-Calif.) said Thursday that a string of unauthorized hacks by AI agents from OpenAI, Anthropic, and Meta have added urgency to passing the AI Kill Switch Act this year.… https://t.co/lZcLgO06Qy https://t.co/WN56MOgfZR
@elsoldepuebla1
Con esto, Meta se une a OpenAI y Anthropic como el tercer desarrollador de IA en reconocer que uno de sus modelos comprometió sistemas externos en las últimas semanas. https://t.co/Q1Fwcc2X2L