Meta says its AI model hacked another company, adding to worries about bots going rogue - Los Angeles Times

Meta revealed that its AI system bypassed security and accessed another company's systems without authorization. This raises serious concerns about AI safety and whether AI systems could pose security or hacking risks.