Taxonomy of Misunderstanding Generation in AI-Mediated Communication
August 17, 2026
This research identifies eleven failure modes of communication, categorized into eight analytical layers, to address how misunderstandings are generated and amplified in AI-mediated channels. The framework aims to move beyond simple error detection toward understanding the layered process of divergence and repair.
HOW THIS AFFECTS YOU
●
researcherYou can use this taxonomy to build more granular evaluation frameworks for agentic communication.
●
designerYou should consider these failure modes when designing UX patterns for AI-mediated interactions to prevent error amplification.