Microsoft research reveals Lazarus backdoor technique that needs no training
What happened
On 11 November 2026, Microsoft's research team disclosed a backdoor technique called Lazarus. According to Microsoft Research (primary source), the technique requires rewriting the weights of fewer than 10 dormant neurons in a large language model and, without any training, lets agents built on that model execute an attacker's commands when a trigger condition occurs. The finding is at the research disclosure stage; the report does not say whether the technique has been exploited in real systems or whether Microsoft has released corresponding defences.
Written by AI from the reports · updated 5 d ago
Timeline
Follow the reports to see the event from each side.
- Microsoft ResearchSelectedLazare, veni foras: Reprogramming Dormant Neurons to Backdoor AI Agents
Microsoft researchers have discovered “Lazarus,” a backdoor technique that can reprogram dormant neurons in an AI model, causing an otherwise normal AI agent to execute malicious commands when a hidden trigger appears. The threat is especially serious for AI agents with tool or terminal access.
Heat of this event
Current heat 3·Peak in the comparable range 9(Nov 11 08:00)·Change over 24 hours in the comparable range -49%
The trend compares only the same accounts observed without a break, so its range can be narrower than the current heat count. Move the pointer or tap the chart to see each hour; with a keyboard, use the left and right arrow keys.