Journal of East China Normal University(Educationa ›› 2026, Vol. 44 ›› Issue (8): 110-126.doi: 10.16382/j.cnki.1000-5560.2026.08.007

Previous Articles    

“The ‘Optimal Learning Style’ Does Not Exist: Large-Scale Learning Analytics Evidence That AI-Driven Personalized Instruction Cannot Outperform Universal High-Quality Teaching” : A Detailed Exposition and a Disclosure Statement of the Research and Writing Process

Jiangshan Sun, Shanshan Chen   

  1. School of Information Technology in Education, South China Normal University, Guangzhou 510631, China
  • Online:2026-08-01 Published:2026-07-22

Abstract:

This paper is one of the publications featured in the “Human-AI Co-Creation Pioneer Papers Ranking”, which is part of the Panoramic Report on the World’s First Large-Scale Social Experiment of “AI as the First Author”. The first part of this paper presents a detailed exposition of the research content: This study draws on large-scale longitudinal data covering 86,237 students and employs double machine learning combined with causal forest algorithms. After controlling for core variables such as students’ previous knowledge and the quality of learning resources, we investigate differences in academic outcomes between AI-customized instruction and universal high-quality instruction. The results show that, after applying double machine learning to control for key confounders, the average treatment effect (ATE) of AI-customized instruction is 0.01 standard deviations, which is neither statistically significant nor reaches the minimum important difference for educational interventions. Heterogeneity analysis further indicates that no specific student subgroups derive additional benefits from personalized paths. The second part is the Transparency Statement for the Research and Manuscript Preparation Process: This study used DeepSeek V3.2 to assist with topic selection, literature review, method design, data analysis, and manuscript writing. Approximately 90% of the initial draft was generated by AI; after systematic revisions by the authors, about 10% of the original AI-generated content remains in the final manuscript. AI is listed as first author to acknowledge its workload, but the core research questions, experimental design, result verification, and final academic judgments were all made by the human authors, who assume full responsibility. All data were de-identified and comply with research ethics. This disclosure aims to accurately document the human–AI collaboration process and to promote discussion on authorship norms.

Key words: AI-customized instruction, universal high-quality instruction, learning style, double machine learning, educational equity