The influence of artificial intelligence (AI) technology on education has been substantial, with generative AI, notably ChatGPT based on large language models (LLMs), standing out as a noteworthy development. ChatGPT’s proficiency in comprehending n...
The influence of artificial intelligence (AI) technology on education has been substantial, with generative AI, notably ChatGPT based on large language models (LLMs), standing out as a noteworthy development. ChatGPT’s proficiency in comprehending natural language semantics and syntax, generating coherent text, and engaging in context-specific dialogues introduces transformative possibilities for education. Its adeptness in intricate interactions with learners suggests that analyzing these interactions can serve as a catalyst for examining individual learning experiences, thereby fostering the evolution of learning analytics. In this context, this research aims to explore the changes in the educational landscape brought about by the adoption of generative AI from a learning analytics perspective and propose directions for its adaptive evolution. To achieve this, we investigated the interaction mechanisms between humans and ChatGPT, seeking to understand the unique characteristics of their engagements. Through this exploration, we identified explanatory possibilities for phenomena that traditional learning analytics could not elucidate and proposed directions for the adaptive evolution of learning analytics, encompassing descriptive, diagnostic, predictive, and prescriptive aspects. This paper is anticipated to contribute to the comprehension of new educational paradigms induced by generative AI, offering insights for more effective, efficient, and engaging instructional design in educational settings where generative AI is integrated.