HumanAI Collaboration in Mathematical Research: A Case Study on an Information-Theoretic Approach to the Riemann Hypothesis

This paper presents a case study of a sustained humanAI collaboration in mathematical research. Over four days in February 2026, the author and Claude (Anthropic, Opus model) jointly developed an information-theoretic framework for the Riemann Hypothesis, producing a preprint with original theorems, analytical proofs, and numerical verications. We document the methodology, analyze the division of cognitive labor, identify the factors that enabled productive collaboration, and discuss the limitations encountered. The principal ending is that the eectiveness of the collaboration derived from complementary strengths human strategic direction and mathematical intuition combined with AI computation, formalization, and iterative capacityrather than from the AI independently generating mathematical insight. We propose that the workow developed here, centered on a cumulative continuation document and rapid explorationformalization cycles, constitutes a replicable methodology for AI-assisted mathematical research.