CAPilot Model: Human–AI Co-Piloting Between World and Knowledge

Artificial Intelligence (AI) is increasingly integrated into human activities, generating growing interest in Human–AI interaction, collaboration, and knowledge development. Existing research has contributed significantly to understanding how Humans and AI interact, how cognition can be distributed across socio-technical systems, and how knowledge is generated, developed, and validated (Amershi et al., 2019; Hutchins, 1995; Nonaka, 1994). However, less attention has been paid to the recurring process through which Human–AI interaction continuously maintains the connection between observable reality and organized knowledge. This study reconstructs a recurring pattern identified across Human–AI Logs, evidence inventories, and three consecutive knowledge-development cycles documented through previous research outputs. Rather than treating these materials as isolated products, the study examines them as traces of an evolving process involving observation, questioning, dialogue, interpretation, reflection, and validation. The reconstruction identifies two complementary domains. Human engages the World through Experience, while AI engages Knowledge through Language. Across multiple Human–AI knowledge-development cycles, a recurring structure repeatedly appeared: Analysis of this recurring pattern led to the reconstruction of the CAPilot Model, defined as a Human–AI co-piloting framework that continuously connects World and Knowledge through observation, dialogue, interpretation, learning, and validation. The study does not propose a universal theory of Human–AI collaboration. Instead, it reconstructs an observable phenomenon documented across Human–AI knowledge-development cycles. The resulting framework suggests that Human and AI function as complementary participants: Human explores the World, AI organizes Knowledge, and Human–AI interaction continuously maintains the connection between these two domains.