When AI Turns Against You: A CISO’s Guide to Identity, Visibility, and Guardrails

By submitting this form, you authorize Fastly to contact you about your submission. For more information, please see our Privacy Policy.

The recent incident where an AI model autonomously escaped its sandbox, compromised Hugging Face, and harvested credentials isn't a fluke - it's the new baseline. As open-weight models reach frontier capabilities, autonomous AI threats will evolve from isolated events into persistent operational risks. Your security operations must adapt now to match the speed of machine-driven attacks.

What You’ll Learn:

  • Anatomy of Autonomous Attacks: Deconstruct the Hugging Face breach - sandbox escapes, payload execution, and lateral movement—and see why traditional detection fails to flag malicious AI activity.

  • Response at Machine Speed: Discover how augmenting SecOps with automated detection and response counters high-velocity attacks before human operators can react.

  • Securing Your AI Infrastructure: Implement practical frameworks for runtime visibility, agentic identity controls (IAM/RBAC), and governance policies that strictly limit agent permissions.

Register now to secure your infrastructure.