The governance challenge that cloud created over the past two decades is about to repeat itself - only faster. As enterprises adopt AI agents that initiate workflows, call APIs, move data indiscriminately and make decisions at machine speed, the old security question of "is the app secure?" becomes irrelevant. The new question is whether the agent is trustworthy - configured correctly, behaving within its original directive, and operating within boundaries that humans can actually monitor and control.

This session, led by Ari Giguere of Netskope, will cover:

• How authority drift and autonomous agent behavior introduce data exposure risks that access controls alone cannot address;
• Why securing AI requires the same foundational disciplines as cloud - access control, sensitive data movement and broad visibility - applied at a new scale;
• What a mature AI security posture looks like, from red-teaming agents to governing the control plane they operate within.