AML.CS0072
AI Recommendation Poisoning via Crafted AI Assistant Links
- Type
- Incident
- When
- 2026-02-10
- Actor, as ATLAS names it
- Multiple commercial entities; 31 distinct companies identified
- Target
- Users of AI assistants with persistent memory features, including Microsoft 365 Copilot, ChatGPT, Claude, Perplexity, and Grok
- Reported by Microsoft Defender Security Research Team.
- On ShieldMap: ChatGPT, Microsoft 365 Copilot (matched by name in the target).
Techniques, in order
- 1Obtain CapabilitiesResource Development · AML.T0016
- 2Stage CapabilitiesResource Development · AML.T0079
- 3Crafted AI Assistant LinksInitial Access · AML.T0131
- 4LLM Prompt ObfuscationDefense Evasion · AML.T0068
- 5AI Agent Context Poisoning: MemoryPersistence · AML.T0080.000
- 6AI Agent Response BiasingImpact · AML.T0130
How these attacks work
Software used
- AI Share URL Creator · Tool
- CiteMET · Tool
From MITRE ATLAS 2026.10 (Apache-2.0). ShieldMap lists who and what was involved and the techniques used, and links to ATLAS for the write-up; it does not copy the steps. Attribution is ATLAS's, not ShieldPi's.