Anthropic Plans Enterprise Safety System Requiring 30-Day Data Retention

Artificial intelligence startup Anthropic is preparing to introduce a updated safety system for its commercial customers later this year, according to a report from Bloomberg's Rachel Metz.

Under the planned framework, corporate clients deploying Anthropic's AI technology will be required to maintain records of user prompt and interaction data for a mandatory 30-day window. The rule is intended to support post-hoc safety audits and monitor potential policy violations across commercial workloads.

To accommodate strict enterprise security requirements, Anthropic plans to allow organizations to store these required 30-day logs directly within their own cloud infrastructure, rather than mandating central storage on Anthropic's servers.

The option to maintain local or private cloud control over compliance data addresses a major friction point for business clients. Large enterprises routinely operate under regulatory frameworks and internal policies that forbid sharing proprietary operational logs with third-party vendors.

By bridging safety auditing requirements with flexible data custody, Anthropic aims to maintain oversight over model misuse while mitigating privacy and security concerns that could otherwise slow institutional adoption.

The new enterprise safety protocols are slated for deployment before the end of the year, with technical implementation guidelines expected closer to the formal launch.