- US AI startup Cathedral breached Anthropic's internal network in under 12 hours during a bug bounty programme, exposing internal Slack messages, staff salaries, and meeting transcripts.
- Anthropic confirmed that user conversations with its Claude chatbot and core AI source code were not compromised, and Cathedral's access was cut off the same day.
- Security experts described the breach as a basic supply-chain hygiene failure involving third-party software rather than a sophisticated attack on AI infrastructure.
- An anonymous Trump administration official warned that the breach demonstrates security vulnerabilities that foreign adversaries like China could exploit.
- The security incident was revealed shortly after Anthropic filed for an initial public offering, which is expected to be the largest stock listing in history.
IN FULL
