5 Biggest Challenges of AI in Cybersecurity

**Executive summary:** This article outlines the five biggest AI-related challenges for cybersecurity—faster time-to-exploit, expanded non-human identities and agent risk, the need for connected visibility across the attack chain, the prevalence of shadow AI, and governance gaps—and recommends identity-centric controls, least-privilege for AI agents, continuous IAM, and using AI-aware testing to reduce risk.