AI agents transform digital authentication models

The rise of AI agents, such as Instinct, Meta Muse, and Grok Bot, is shifting the traditional authentication model. Unlike standard user-driven logins, these agents perform tasks like booking flights, managing subscriptions, and accessing healthcare portals autonomously, often while the user is inactive.

To function, these agents require the ability to inherit a user’s authenticated identity. Instinct has addressed this through a feature called ‘Vault,’ where users store credentials. According to founder Noah Shinn, 32% of Instinct users were sharing credentials via Vault by early September, with 37% storing at least one account password within their first three weeks. The platform has also added support for TOTP authenticators, allowing the agent to generate its own six-digit verification codes.

Securing these platforms requires a comprehensive authentication subsystem. This includes managing various identity types, such as human users, service accounts, and machine credentials, while maintaining a strict distinction between authentication (verifying identity) and authorization (determining permissions).