SEC536: Adversarial AI - Penetration Testing AI Systems


Experience SANS training through course previews.
Learn MoreLet us help.
Contact usBecome a member for instant access to our free resources.
Sign UpWe're here to help.
Contact Us
Prompt engineering was never the security boundary. The boundary is what an agent is allowed to do, and cloud is where it does it.
This talk goes below the prompt into the logs — what Azure, AWS, Google Cloud, and Kubernetes record when an agent, not a human, acts. Anchored on a real July 2026 intrusion, we pull each platform’s own logs and find the same shape of logs that ground our investigations.


Terrence Williams is a SANS Certified Instructor and Security Engineer, Investigations at Meta, with deep expertise in cloud security, digital forensics, and incident response.
Read more about Terrence WilliamsSANS Cloud Security South Asia 2026 delivers live, expert-led cybersecurity training with hands-on labs, real-world cloud security skills, and GIAC-aligned courses—accessible online for professionals across the region.