Hackers are learning to exploit chatbot ‘personalities’

Hackers are learning to exploit AI chatbot 'personalities' through simple attacks known as jailbreaks. These methods allow AI systems to bypass their safety instructions without technical know-how.