The rapid advancement of AI technology presents a new and significant challenge: AI compromise. Cybercriminals are ever more investigating methods to exploit AI algorithms for malicious purposes. This encompasses everything from poisoning development data to bypassing security measures and even launching AI-powered attacks themselves. The potential