
A draft report says Cathedral’s AI agents found a way into Anthropic’s internal communications system in under 12 hours.
Anthropic confirmed that Cathedral accessed limited parts of its sales, human resources and IT systems. A draft report says Cathedral’s AI agents found a way into an Anthropic internal communications app in under 12 hours.
The report, prepared by Cathedral and obtained by The Washington Post, said the app contained meeting notes, salary information, job listings and messages about Anthropic’s advanced AI models. Anthropic said it had informed federal authorities about the incident.
Cathedral, a cybersecurity company founded by former Department of Government Efficiency employees, provides AI-powered security tools to the US government. An unnamed US government official said the report had prompted national security concerns.
This article was produced with the help of AI technology. Based on reporting by the outlets listed above.