Microsoft blocks Claude Fable 5 over data fears
microsoft banned internal use of anthropic's claude fable 5 due to its 30-day data retention policy, signaling escalating tensions between enterprise data security and ai safety requirements.
microsoft has blocked internal access to anthropic's new claude fable 5 model, while still allowing other claude models, because fable 5 requires storing prompts and outputs for 30 days to function. this data retention policy is necessary for anthropic's new safety classifiers but violates microsoft's strict zero data retention rules for handling confidential information. legal teams are assessing whether the risk of retaining data, which extends to two years for policy violations, is acceptable for internal corporate use. this conflict illustrates a widening gap between the data demands required for cutting-edge ai safety mechanisms and the strict data privacy mandates of large enterprises.