Large language model-assisted modern programming pedagogy model: a perspective on teaching programming in the era of LLMs

Abstract

Large language models (LLMs) are a double-edged technology. They can serve as learning assistants in programming while simultaneously raising pedagogical and ethical concerns. Programming education researchers have responded by providing guidelines and recommendations on teaching, assessment, and the ethical use of LLMs. Although these developments are welcome, the literature still lacks theory-based frameworks that provide a philosophical grounding for programming pedagogy. This perspective paper addresses this gap by proposing the LLM-Assisted Modern Programming Pedagogy (LAMPP) model. The model aims to provide a theory-informed framework for systematically guiding the integration of LLMs into programming pedagogy. It proposes four design principles encompassing the core components of pedagogy (i.e., teaching, lesson design, engagement, and assessment). Ethics is included as an additional dimension because ethical concerns cut across all aspects of pedagogy. The LAMPP model proposes different mechanisms that may mitigate potential learning losses associated with inappropriate or excessive reliance on LLMs. Future studies should test the effectiveness of these mechanisms. The model may provide a basis for educators, researchers, and policymakers when considering policies, curriculum structures, course outcomes, assessment practices, and ethical guidelines related to programming pedagogy. In summary, it is not about LLM tools themselves but about the pedagogical orchestration of these technologies in programming education.

The authors' abstract, as published at the source. Frontiers in Artificial Intelligence, 2026 · DOI ↗