Rethinking Human-AI Collaboration in Information Seeking: Why Epistemic Incompatibility Demands New Design Paradigms

The integration of artificial intelligence agents into information retrieval systems has prompted two dominant narratives: AI as replacement for human information seeking, and AI as collaborative partner. This perspective paper challenges both framings by proposing a third alternative – that human information seeking behavior and AI agent functionality represent fundamentally incompatible epistemological paradigms. Drawing on information behavior theory and critical AI scholarship, we argue that this incompatibility stems from three dimensions: epistemic orientation (learning vs. pattern reproduction), temporal structure (processual vs. instantaneous), and agentic purpose (uncertainty resolution vs. task execution). Rather than forcing integration through collaboration metaphors or accepting wholesale replacement, we advocate for system designs that acknowledge and preserve these fundamental differences. This reframing has profound implications for the design of information interaction systems, evaluation methodologies, and research directions in human information interaction and retrieval.