Home / Technology / Okta CEO: AI agents need new identity layer
Okta CEO: AI agents need new identity layer
30 Mar
Summary
- AI agents require a new identity layer, distinct from human users.
- Okta is developing a 'kill switch' for rogue AI agents.
- Agent identity is a hybrid between human and system attributes.
Okta CEO Todd McKinnon discussed the evolving landscape of enterprise software in the age of AI, acknowledging a 'Saaspocalypse' threat but emphasizing a significant opportunity in identity management for AI agents. He described agent identity as a hybrid between human and system attributes, necessitating a new layer of security and control.
McKinnon detailed Okta's strategy to become the identity framework for AI agents, which could represent a massive new category in cybersecurity. He highlighted the need for a 'kill switch' to manage rogue agents and stressed that controlling these agents' access to data and systems is paramount.
Okta's blueprint for the agentic enterprise focuses on onboarding agents with unique identities, standardizing their connection points, and providing mechanisms to revoke access if agents behave unexpectedly. This proactive approach aims to balance the flexibility of AI agents with necessary security controls.
