Anthropic Introduces Enterprise Frontier Safeguards for AI Misuse Detection

Anthropic is announcing Enterprise Frontier Safeguards (EFS), a new solution that integrates zero data retention (ZDR) privacy with advanced systems for detecting AI misuse. EFS stores customer data within their own controlled cloud infrastructure, not Anthropic's, and will be rolled out in phases to customers starting later this fall.

EFS is designed to mitigate the potential for misuse and autonomous misbehavior from Mythos-class models like Claude Fable 5.1, including sophisticated cyberattacks and credential theft. It allows for the necessary cross-session and cross-account monitoring for effective threat detection while respecting enterprise data retention policies.

With EFS, customers maintain control over how data is reviewed and where it is stored. Misuse signals from automated monitoring systems are sent directly to customers for internal review, and activity data can reside in the customer's own cloud account under their encryption keys and access policies.

The system provides automated safety monitoring without requiring human review by Anthropic employees. This design protects sensitive information, allowing customers in regulated industries to deploy frontier models safely by ensuring their teams handle any flagged data.