09/10/2026
AI security is no longer just about preventing harmful outputs.
As AI agents gain access to tools, data, memory, and production systems, the bigger question becomes: what can they actually do?
In this episode of the ODSC Ai X Podcast, host Dan Gerlanc speaks with David Campbell, previously Head of AI Security at Scale AI and a founding member of the OWASP AIVSS Project and AIUC-1, about how AI security is changing as agents become more capable.
David brings nearly 20 years of engineering and security experience across Scale AI, Uber, DoorDash, Google, and Nest, with deep work in large-scale AI red teaming, resilience, and production security.
The conversation explores why securing the model alone is not enough, how persistent prompt injection creates new risks, and why observability, least privilege, and system-level controls are becoming essential for agentic AI.
Topics include:
- Why AI security must operate at the systems layer
- How agent access changes the risk model
- Why human red teaming still matters
- Persistent prompt injection and why “all context is vulnerable”
- Least privilege for AI agents
- Logging, telemetry, observability, and replaying agent actions
- Why autonomous cyber defense may become necessary
As David puts it:
“Security happens at the systems layer.”
🎧 Listen here: https://linktr.ee/odsc
