Anthropic Unveils Enterprise Frontier Safeguards for Regulated Firms
Anthropic's Enterprise Frontier Safeguards addresses concerns from regulated firms about data storage and review. The system allows customers to store their logs in their own cloud accounts, under their encryption keys and access policies, with automated review that doesn't involve human intervention from Anthropic employees.
The company worked with the Analysis and Resilience Center for Systemic Risk, a group of CISOs from major banks, to develop the solution. The goal was to allow regulated firms to use Anthropic's models while meeting their obligations to customers, employees, and regulators.
Anthropic has dropped zero data retention in favor of a 30-day retention window for its most capable tier. This change allows for detection of sophisticated misuse that may not be caught by scanning individual interactions.
The system will support multiple platforms, including Claude Code, Claude Enterprise, and Amazon Bedrock, with customer-owned storage and encryption keys. Rollout is planned in phases, with broad availability targeted for later this fall.