Security researchers from the emerging cybersecurity start-up Cathedral successfully breached internal systems at artificial intelligence pioneer Anthropic. The unauthorized access exposed potential vulnerabilities in internal developer tooling and corporate infrastructure, raising immediate security alarms. Cathedral disclosed the findings to Anthropic’s security team under responsible disclosure protocols, allowing the AI maker to patch the exposed endpoints. The incident underscores the severe cybersecurity challenges facing foundational model labs as they handle sensitive proprietary data and rapid product scaling. Industry experts emphasize the critical need for rigorous offensive security testing and robust defense-in-depth measures across the generative AI sector.

Full exclusive : How an AI start-up Cathedral broke into Anthropic’s internal systems.