2026-11-11 · Help Net Security · MEDIUM

Microsoft researchers demonstrate hidden backdoor technique for AI agents

60-WORD SECURITY BRIEF

Microsoft researchers have demonstrated how small changes to an AI model can implant a hidden backdoor, causing an agent to execute an attacker’s command when a specific trigger appears. The technique, called Lazarus, requires no model retraining and allows the agent to behave normally during standard safety evaluations.