Due to concerns about data privacy and information security, Microsoft has recently implemented restrictions within the company. According to tech media reports, Microsoft has banned its employees from using the latest top-tier large model, Claude Fable 5, released by the AI startup Anthropic. The core reason for this restriction lies in the model's new data retention policy.
Touching the Data Retention Red Line
Although Microsoft quickly launched the model on its platform for external paying customers, it has removed it from the list of internal tools for employees. Other models in the Claude series, which strictly adhere to the "zero data retention" rule, can still be used internally. The Microsoft legal team is conducting a comprehensive assessment of this controversial new rule, and the model cannot obtain internal approval before that.
Controversy Caused by Security Defenses
The data retention policy that sparked the controversy was actually established by Anthropic to support its new security classification system. Due to Claude Fable 5's strong capabilities in sensitive tasks such as cybersecurity, the company added protective mechanisms, resulting in user questions and outputs being retained for 30 days, with违规 data possibly stored for up to two years. This clause, which poses a potential risk to corporate core assets, ultimately triggered Microsoft's compliance concerns.
