Composition writing is an essential component of language learning, but grading compositions can often be time-consuming and uneven. Teachers grapple with inconsistent grading standards, and the sheer quantity of time spent grading prevents them from offering effective, interactive feedback. These are especially salient issues in Chinese language instruction, where automated essay scoring has not been as well developed as it has been in English-language education.
To overcome these issues, a group of researchers led by Chanjin Zheng and Wei Xia from East China Normal University and Shaoguang Mao and Yan Xia from Microsoft Research Asia has created the ELion Intelligent Chinese Composition Tutoring System, an artificial intelligence-based system that aims to improve essay assessment. Their research, published online in ECNU Review of Education on February 4, 2025, discusses the development of the system and its influence on Chinese language education.
"We've seen an amazing shift in how AI can assist Chinese language learning," Zheng et al. continue. "The transition from BERT to the combination of ChatGPT and beyond shows the power of integrating various AI technologies to develop holistic learning solutions."
In the research, the authors describe how the ELion system works by assessing student essays along four major dimensions: topic understanding, content, expression (fluency, logic, and grammar), and handwriting. Employing the Bidirectional Encoder Representations from Transformers (BERT) model at first, the system assessed the linguistic attributes of student essays and gave them thorough, unbiased ratings. This model was enhanced further by the inclusion of ChatGPT, allowing the system to provide more subtle and personalized feedback, hence promoting student creativity and engagement.
The incorporation of ChatGPT with BERT has dramatically enhanced the delivery of feedback. The system not only provides thorough critiques of students' writing but also promotes an interactive conversation between the AI and students, facilitating ongoing improvement in their writing capabilities.
The research points out that the ELion system has played a crucial role in solving the long-standing problem of the overburdened grading work for teachers, especially in Chinese primary and secondary schools. Since its establishment in spring 2021, the system has grown at an incredible pace, with 251 institutions of learning having embraced the technology in China as of June 2024. With more than 15,000 students and 560 teachers utilizing the platform, the system has enabled the grading of close to 50,000 compositions.
A teacher at a Shanghai primary school highlighted the importance of the ELion system in class: "ELion has freed Chinese language teachers from the drudgery of essay correction, enabling them to spend more time on interactive learning and developing creativity." The system has also reduced the teachers' workload, releasing time for more concentrated, creative teaching.
Despite its success, the researchers acknowledge that the widespread implementation of AI in education presents challenges. While technologies like ELion have shown strong potential to improve grading efficiency and deliver personalized feedback, it remains essential to integrate these tools thoughtfully into classroom settings. The real value of AI lies in its ability to support not replace the human aspects of teaching.
The research also covers the need to continue developing AI systems to be aligned with various educational requirements. BERT and ChatGPT integration into the ELion system provides an example of how AI can be specialized to serve different educational environments. The authors emphasize that the capacity to integrate old and new technologies lies in making sure that AI-based tools have the ability to make significant contributions to the educational environment.
As the ELion system continues to reach further and mature its capabilities, the researchers note that AI in learning must remain adaptive and flexible to both the needs of teachers and learners. They note that through the use of a combination of AI tools, the environment can become even more personalized, efficient, and engaging, preparing students for future challenges.
This research provides an important insight into the potential of AI in restructuring education, drawing attention to the role of modern technologies in personalized learning and in easing teachers' workload. Continuous development of the ELion system serves as an example for coming innovations in educational tools based on AI.
