Amazon: AI agents exploit a zero-day flaw in eleven minutes
Amazon Threat Intelligence says a suspected China-linked actor used the Cairn agentic AI framework to gain root access in eleven minutes.
The path-traversal flaw affected MONAI Label, a medical-imaging application, on an East Asian university server. Amazon disclosed it on November 20. Sources disagree on whether a patch exists and on the confidence level of the attribution. Amazon's CISO CJ Moses reportedly said autonomous AI exploits bugs faster than most security teams can triage alerts.
Source: AI Weekly (aiweekly.co), published on November 21, 2026.