Anthropic Lets Enterprises Keep Claude Misuse-Detection Logs in Their Own Cloud

Summary
Anthropic announced safeguards that let enterprise customers store Claude misuse-detection logs in their own cloud under their controls. Automated alerts go to customers, not Anthropic staff; broad availability is targeted for later this fall.
Key points
- Enterprise Frontier Safeguards lets customers store misuse-detection activity data in Amazon S3, Azure Blob Storage, or Google Cloud Storage under their own encryption keys, access policies, and audit logging.
- When automated systems flag activity, alerts go to the customer; Anthropic says its employees do not review the flagged data.
- Monitoring targets serious misuse, including offensive cyber or biological activity, stolen or leaked credentials, and destructive behavior by agents.
- Anthropic says detecting patterns across sessions and accounts requires retaining data for a meaningful period; its most capable tier moved from zero data retention to a 30-day window starting with Fable 5.
- The monitoring window’s duration is not disclosed. Eligible customers will receive zero data retention on Fable 5 and Fable 5.1 until safeguards reach them.
- Support is planned across Claude products and cloud platforms, with broad availability targeted for later this fall. The controls are opt-in and carry no Anthropic charge; customers pay their cloud provider for storage and data use.
Article Details
- Event Type
- Product and security-control announcement
- Impact
- Anthropic introduced Enterprise Frontier Safeguards, allowing eligible customers to keep misuse-detection logs in customer-controlled cloud storage under their own encryption keys and access controls. Automated alerts go to the customer, and Anthropic employees do not review flagged data. Rollout is planned in phases, with broad availability targeted for later in the fall; Anthropic says there is no added charge for the safeguards.
People
Munish Kumar SharmaCISO at Wells Fargo; said the safeguards let the bank retain custody of its logs and keys while Anthropic operates detection.Scott DePasqualePresident and chief executive of the Analysis and Resilience Center for Systemic Risk; described the requirements its members set for using frontier models in systemically important banks.
Vendors
AmazonActivity data used for misuse detection can live in the customer’s Amazon S3, Azure Blob Storage, or Google Cloud Storage account, under the customer’s encryption keys, access policies, and audit logging.AnthropicAnthropic introduced Enterprise Frontier Safeguards to let regulated customers keep misuse-detection logs in their own cloud storage under their own encryption keys and access controls.GoogleActivity data used for misuse detection can live in the customer’s Amazon S3, Azure Blob Storage, or Google Cloud Storage account, under the customer’s encryption keys, access policies, and audit logging.MicrosoftEnterprise, the Claude Platform, Amazon Bedrock, Claude Platform on AWS, Google’s Agent Platform, and Microsoft Foundry, with equivalent controls whether you buy direct or through a cloud partner.
Products
Amazon BedrockSupport is planned for Claude Code, Claude Enterprise, the Claude Platform, Amazon Bedrock, Claude Platform on AWS, Google’s Agent Platform, and Microsoft Foundry, with equivalent controls whether you buy direct orClaudeThe system routes automated alerts to customers instead of Anthropic staff, aiming to support safer use of Claude across major banks and enterprises while preserving data custody.Claude CodeSupport is planned for Claude Code, Claude Enterprise, the Claude Platform, Amazon Bedrock, Claude Platform on AWS, Google’s Agent Platform, and Microsoft Foundry, with equivalent controls whether you buy direct orClaude EnterpriseSupport is planned for Claude Code, Claude Enterprise, the Claude Platform, Amazon Bedrock, Claude Platform on AWS, Google’s Agent Platform, and Microsoft Foundry, with equivalent controls whether you buy direct orClaude PlatformSupport is planned for Claude Code, Claude Enterprise, the Claude Platform, Amazon Bedrock, Claude Platform on AWS, Google’s Agent Platform, and Microsoft Foundry, with equivalent controls whether you buy direct orEnterprise Frontier SafeguardsAnthropic introduced Enterprise Frontier Safeguards to let regulated customers keep misuse-detection logs in their own cloud storage under their own encryption keys and access controls.Fable 5Anthropic dropped it for its most capable tier starting with Fable 5 and moved to a 30-day retention window.Fable 5.1Rollout happens in phases with broad availability targeted for later this fall, and eligible customers get zero data retention on Fable 5 and Fable 5.1 until their turn comes.Google’s Agent Platformfor Claude Code, Claude Enterprise, the Claude Platform, Amazon Bedrock, Claude Platform on AWS, Google’s Agent Platform, and Microsoft Foundry, with equivalent controls whether you buy direct or through a cloudMicrosoft Foundrythe Claude Platform, Amazon Bedrock, Claude Platform on AWS, Google’s Agent Platform, and Microsoft Foundry, with equivalent controls whether you buy direct or through a cloud partner.
