Anthropic announced on Tuesday the launch of Enterprise Frontier Safeguards, a new system designed to strengthen privacy and detect potential misuse of its Claude AI models.
The system combines zero data retention capabilities with automated safeguards that can identify suspicious activity.
Customers Can Keep Data in Their Own Cloud
One of the main features of Enterprise Frontier Safeguards is customer-controlled data storage.
Instead of storing information on Anthropic’s systems, customers can keep their data within their own cloud infrastructure.
Supported storage options include Amazon S3, Microsoft Azure Blob Storage and Google Cloud Storage.
Anthropic said the feature will be introduced gradually, with the rollout expected to begin later this fall.
Anthropic Worked With More Than 100 Customers
The company developed Enterprise Frontier Safeguards in collaboration with more than 100 customers.
These organizations operate across industries including financial services, healthcare, manufacturing, telecommunications, law, retail and the public sector.
Anthropic also worked with major cloud providers, including Amazon Web Services, Google Cloud and Microsoft Azure.
The system is expected to be available across several Claude products and platforms.
These include Claude Code, Claude Enterprise, the Claude Platform, Amazon Bedrock, Claude Platform on AWS, Google’s Agent Platform and Microsoft Foundry.
Zero Data Retention Expands Claude Privacy
Anthropic said eligible customers will receive zero data retention for Fable 5 and Fable 5.1 until Enterprise Frontier Safeguards becomes widely available.
The company previously introduced a 30-day data retention period with Fable 5.
According to Anthropic, the change was designed to help address attempted misuse of AI systems, including fraud and cyberattacks.
The company also emphasized that it has never used enterprise customer data for model training without permission.
Financial and Cybersecurity Leaders Helped Shape the System
Anthropic worked with the Analysis and Resilience Center for Systemic Risk while developing the safeguards.
The organization includes chief information security officers from major financial institutions such as Goldman Sachs, Morgan Stanley, Citi, Bank of America and Wells Fargo.
Anthropic also consulted with executives and security leaders from Comcast, KPMG, Mastercard, Salesforce and Visa.
Their involvement helped shape the system’s approach to enterprise security and risk management.
Automated Systems Monitor Potential Misuse
Enterprise Frontier Safeguards uses automated systems to analyze traffic for signs of misuse.
The system can detect suspicious activity related to offensive cyber capabilities, biological threats and potentially stolen credentials.
When monitoring identifies patterns that require further attention, the system sends signals directly to the customer for review.
Anthropic said its own employees do not need to conduct human reviews of the customer data.
Enterprise Safeguards Remain Optional
Several features included in the system are opt-in.
These include customer-owned storage, Customer-Managed Encryption Keys and automated traffic review.
Anthropic said enabling these features will not change Claude’s model behavior, API pricing or rate limits.
The company also does not charge an additional fee for Enterprise Frontier Safeguards.
However, customers who choose to store data in their own cloud infrastructure will still be responsible for storage and related charges from their cloud provider.
The new safeguards reflect Anthropic’s broader effort to strengthen Claude’s appeal to enterprise customers while balancing privacy, security and AI misuse prevention.
