The quick advancement of AI technology presents the novel and significant challenge: AI compromise. Cybercriminals are increasingly exploring methods to abuse AI systems for harmful purposes. This involves everything from tampering learning data to evading security safeguards and even launching AI-powered attacks themselves. The potential effects o