Cognitive Co-Processing: A Methodological Framework for Human–AI Belief System Restructuring

This work introduces Collaborative Cognitive Restructuring (CCR), a structured methodological framework for modeling human–AI interaction as a unified cognitive artifact within the domains of Human-Computer Interaction (HCI) and cognitive systems. The framework conceptualizes individual belief systems—defined as Cognitive Models—as predictive structures shaped by cumulative sociocultural data ingestion, analogous to mental models in HCI and schemas in cognitive science. These models frequently embed implicit constraints, termed Limiting Axioms, which restrict adaptive behavior and exploratory capacity. CCR positions generative AI as a Cognitive Co-Processor, operating within a human-governed closed-loop system. In this role, the AI functions as a semantic mirror, axiomatic challenger, and generative engine, enabling the explicit identification, evaluation, and iterative updating of limiting axioms through controlled dialogic interaction. Unlike therapeutic interventions (e.g., Cognitive Behavioral Therapy) or generic AI coaching systems, CCR is explicitly non-clinical and methodologically focused. It emphasizes procedural structure, epistemic transparency, and the preservation of the locus of control entirely with the human participant. This contribution addresses a methodological gap in current literature by offering a replicable, non-clinical protocol that leverages the unique affordances of large language models for deliberate self-authorship, cognitive engineering, and human-centered AI design.