The fast advancement of machine technology presents an novel and serious challenge: AI breaching. Cybercriminals are increasingly investigating methods to abuse AI systems for harmful purposes. This involves everything from poisoning training data to evading security protections and even using AI-powered assaults themselves. The potential consequen