RunSybil CEO Ari Herbert-Voss joins the Google Cloud Security Podcast to discuss whether AI-powered red teaming is genuinely novel or just automated known TTPs, how RunSybil turns AI findings into actionable, fixable tickets, the guardrails that keep an autonomous "hacker AI" safe in production, and why authentication and authorization bugs are where the technology shines. The conversation closes with a case study: an AI-discovered remote code execution vulnerability that had gone unnoticed by human researchers for years.