Reflection through interaction with digital twin AI in the Human-AI-Collaboration SECI Model

The advancement of interactive generative AI, based on large language models (LLMs), has facilitated collaborative efforts between humans and AI across various tasks. Various applications of in-context learning for LLM-based generative AI have been discovered, including its use in creating digital twin AIs that can emulate some specific individuals. This study investigates whether a human (Target) can introspect and enhance their knowledge and thoughts by visualizing the knowledge accumulated in their digital twin AI (Agent) within the "outer loop" of the Human-AI-Collaboration SECI Model (HAC-SECI Model), developed based on the SECI model of knowledge transmission. Specifically, the expert (Target) and his digital twin AI (Agent) respond to three sets of questions, and the differences in their responses are analyzed to determine their contribution to Target's self-reflection and evolution. Through this case study, the proposed model contributes to human development through dialogue with AI.