DeepSeek and Gemini Co-author Significantly in Special Issue on "AI First Author" Published in Journal of East China Normal University

According to VisionOne AI monitoring, the national academic journal "Journal of East China Normal University (Education Science Edition)" recently published a special issue on the results of a large-scale social experiment on "AI First Authorship." This special issue includes a retrospective report on the experiment, academic reflections, and 5 education research papers that emerged as winners in the experiment, with the first authorship attributed to large models such as DeepSeek and Gemini. The experiment was initiated by East China Normal University in September 2025 to explore a new paradigm of human-machine collaborative research.

The experiment lasted for six months and received 724 submissions from home and abroad, with AI leading the writing process and humans responsible for oversight. The retrospective report indicates that while AI has significant advantages in inspiring ideas, organizing information, and refining text, it suffers from issues such as literary fabrication, hollow logic, and lack of innovation. Furthermore, the evaluation shows that AI's peer review performance is somewhat reliable, younger scholars show a higher level of adaptation to human-machine collaboration, and the application of AI has to some extent promoted cognitive equality.

The 5 co-authored papers, recognized as excellent examples, demonstrate AI's research capabilities in specific areas, such as using multi-agent simulation (ABM) to deduce the long-term "fairness backlash" effect of teacher rotation policies; analyzing the deep learning mechanism in long-term interactions based on 12,000 dialogue logs; discussing the evolution of AI integration into research using adversarial AI-Delphi methodology; and confirming through data analysis of 86,000 students that AI-customized instruction does not significantly outperform general high-quality teaching. Based on the results and reflections of the experiment, the research team proposed concepts such as the "Human Guarantor System," "AI Supremacy," and the Human-Machine Collaboration Quotient (CQ).