Anthropic accidentally exposed unreleased AI models and confidential executive information through an unsecured content management system

Details of unreleased AI models were inadvertently made publicly accessible via Anthropic's content management system

Confidential materials including unpublished blog drafts and private company information were exposed in the security lapse

The incident affected sensitive data related to an invite-only CEO retreat

This represents a significant security oversight by the AI safety-focused company