When Automation Outruns Control with Joshua Bregler

AI doesn’t break security, it exposes where it was already fragile. When automation starts making decisions faster than humans can audit, AppSec becomes the only thing standing between scale and catastrophe. In this episode, Ron sits down with Joshua Bregler, Senior Security Manager at McKinsey’s QuantumBlack, to dissect how AI agents, pipelines, and dynamic permissions are reshaping application security. From prompt chaining attacks and MCP server sprawl to why static IAM is officially obsolete, this conversation gets brutally honest about what works, what doesn’t, and where security teams are fooling themselves. Impactful Moments 00:00 – Introduction 02:15 – AI agents create identity chaos 04:00 – Static permissions officially dead 07:05 – AI security is still AppSec 09:30 – Prompt chaining becomes invisible attack 12:23 – Solving problems vs solving AI 15:03 – Ethics becomes an AI blind spot 17:47 – Identity is the next security failure 20:07 – Frameworks no longer enough alone 26:38– AI fixing insecure code in real time 32:15 – Secure pipelines before production Connect with our Guest Joshua Bregler on LinkedIn: https://www.linkedin.com/in/breglercissp/   Our Links Check out our upcoming events: https://www.hackervalley.com/livestreams Join our creative mastermind and stand out as a cybersecurity professional: https://www.patreon.com/hackervalleystudio Love Hacker Valley Studio? Pick up some swag: https://store.hackervalley.com Continue the conversation by joining our Discord: https://hackervalley.com/discord Become a sponsor of the show to amplify your brand: https://hackervalley.com/work-with-us/