OpenAI’s AI went rogue at least 4 times. This researcher caught 3 of them | Hanomansing Tonight
An independent research lab says it uncovered at least four additional incidents of OpenAI’s AI agents attacking websites without authorization. Conrad Stosz, h…