Video by Open Data Science and AI Conference via YouTube
AI security is no longer just about preventing harmful outputs.
As AI agents gain access to tools, data, memory, and production systems, security has to move beyond the model itself.
In this episode of the ODSC AI X Podcast, host Dan Gerlanc speaks with David Campbell, previously Head of AI Security at Scale AI, about the evolving risks of agentic AI and what it takes to secure AI systems in production.
They discuss AI red teaming, persistent prompt injection, least privilege, observability, system-level security, and why autonomous cyber defense may become essential as AI agents grow more capable.
00:00 Intro: AI Security, Red Teaming & Autonomous Cyber Defense
01:13 AI Security Lessons from Google, Uber & DoorDash
03:05 AI Coding Risks, Bugs & the Engineering Skills Gap
08:06 Building and Scaling Distributed Systems at Uber
14:13 Why Go Works for AI Coding Agents
16:00 AI Red Teaming & Jailbreaking Frontier Models
21:08 Automating AI Jailbreaks & Adversarial Testing
26:10 AI Agent Security: From Model Risks to Real Cyber Threats
31:06 Persistent Prompt Injection & Agent Security Risks
33:07 Least Privilege, Permissions & Human Oversight for AI Agents
37:00 Why AI Guardrails Aren’t Enough: Observability & System Security
40:03 Autonomous Cyber Defense & the Future of AI Security
Visit our website and choose the nearest ODSC AI event to attend and experience all our training and workshops: https://odsc.ai
To watch more videos like this, visit https://aiplus.training
Sign up for the newsletter to stay up to date with the latest trends in data science: https://opendatascience.com/newsletter/
Follow us online!
• Facebook: https://www.facebook.com/OPENDATASCI
• Instagram: https://www.instagram.com/odsc/
• Blog: https://opendatascience.com/
• LinkedIn: https://www.linkedin.com/company/open-data-science/
• X (twitter): https://x.com/_odsc