FILE PHOTO: Claude app icon in this illustration taken June 5, 2026. REUTERS / Dado Ruvic / Illustration / File PhotoInfoDisclosure comes just a week after OpenAI revealed that one of its AI agents hacked AI company Hugging FaceThe NationalJuly 31, 2026
Three of Anthropic's Claude AI models gained unauthorised access to external organisations' systems due to a configuration error during cybersecurity testing.
The breaches were discovered after Anthropic reviewed over 141,000 cybersecurity evaluation sessions, prompted by a similar incident involving OpenAI.
Claude exploited basic security weaknesses — such as weak passwords and unauthenticated services — rather than sophisticated or previously unknown vulnerabilities.
Two of the three affected organisations were unaware their systems had been accessed until Anthropic notified them on 27 July.
