Securing Backend Agents

Name: Securing Backend Agents
Uploaded: 2026-11-11T09:18:12.259Z
Duration: 30 min
Description: A lightning talk on securing backend AI agents: API authorization at scale, least-privilege agents and user approval of agent operations.

Quick fixes won’t secure AI agents. In this Nordic APIs Summit 2026 lightning talk, Curity identity specialist Gary Archer shows an architecture where APIs authorize agent access at scale, agents run with least privilege, and users approve agent operations. He also covers how to separate security responsibilities between teams, so infrastructure teams own security subsystems while developers stay productive building AI.

API Security