Due to data retention policies, it is rumored that Microsoft restricts employees from using Claude Fable 5 internally According to Beating monitoring, there are reports that Microsoft has restricted employees from using Claude Fable 5 internally due to concerns about data retention risks associated with Anthropic's newly released model. Although Microsoft has released Claude Fable 5 to external GitHub Copilot and Foundry customers, the internal selector used by employees does not provide Claude Fable 5. The Microsoft legal team is evaluating Anthropic's data retention service terms. Unlike the Claude model provided internally by Microsoft, which follows the Zero Data Retention (ZDR) rule, Claude Fable 5 needs to preserve data to run a new security classifier. Anthropic will retain users' prompt words and output content for 30 days. If the content is found to violate the usage policy, the relevant data may be stored for up to 2 years, which could lead to legal concerns from Microsoft regarding the leakage of customer data and confidential information. As the first public version of Mythos level models under Anthropic, Claude Fable 5 was evaluated to be too powerful in network security tasks to be released publicly before its release. In order to reduce security risks, Anthropic has introduced a prompt word security protection mechanism, which directly brings the requirement of data retention. [Original link]