AI exploitation is the use of artificial intelligence tools, models, or generated outputs to support attack activity. That can include writing convincing phishing messages, generating malicious code, automating reconnaissance, translating lures, or producing realistic voice and text impersonation. The term does not mean AI is always the main weapon; it often acts as an accelerator for existing tactics.
It matters because AI can reduce the skill, time, and cost needed to run scams and intrusion campaigns at scale. Defenders may see faster, more polished social engineering, more targeted impersonation, and more adaptive messaging that evades simple detection rules. In practice, security teams respond by hardening identity checks, training users to verify requests out of band, watching for unusual automation patterns, and reviewing how sensitive data is exposed to AI systems that could be abused.