As AI agents become more autonomous—and increasingly capable of discovering APIs, calling tools, accessing enterprise data, and taking actions on behalf of users—the traditional API security model is being challenged in entirely new ways.

This panel brings together security, API, and technology leaders to explore what happens when APIs are no longer primarily consumed by predictable applications and human developers, but by AI agents capable of dynamically deciding what to access, when to act, and how to use the information they receive.

Panelists will examine emerging attack surfaces created by agentic AI, including prompt injection, excessive permissions, identity and authorization abuse, insecure tool access, sensitive data leakage, compromised agents, and unintended API actions. The conversation will also explore how existing approaches to authentication, authorization, API gateways, rate limiting, monitoring, and zero-trust security need to evolve for an agent-driven ecosystem.

Beyond identifying the risks, the discussion will focus on practical strategies organizations can implement today. How should companies authenticate and authorize autonomous agents? How much access should an agent receive? How can security teams monitor agent behavior and identify suspicious or unintended actions? And how do organizations introduce stronger controls without creating friction that slows developers and AI innovation?

Attendees will leave with a clearer understanding of the evolving API security landscape and practical considerations for building secure, observable, and scalable API infrastructure as AI agents become increasingly integrated into enterprise systems.