Coffee, Chaos and ProdSec Ep 23 - Part 2 - AI Security Incident Response, Supply Chain Chaos, AI Training and the Compliance Wake Up Call
Feb 4, 2026
They dig into why traditional incident response and SOC playbooks fail when AI acts unpredictably. They explore prompt injection, data leakage, and gaps in visibility and DLP. They debate automated containment, managing agent blast radius, and risks from vendors enabling AI by default. They also cover AI supply chain questions, the limits of AIBOM ideas, and the need for tailored AI training and compliance readiness.
AI Snips
Chapters
Transcript
Episode notes
Visibility Is Nonnegotiable
- Visibility is the foundation because you can't secure or respond to what you can't see.
- Network bits alone don't answer if sensitive data was copied into public AIs.
Detect Data Exfiltration To Public AI
- Enforce visibility and DLP controls to detect users copying sensitive docs into public AI tools.
- Combine identity monitoring with content-aware detection rather than only network indicators.
Require Sandboxing And Credential Controls
- Ask product teams to design AI with sandboxing, kill switches, monitoring, and credential controls.
- Require these controls before massâdeploying agents or enabling internet access.