Top university limits use of AI tools

Fudan University, a top-notch university in Shanghai, has recently issued a regulation on the use of AI tools in writing undergraduate graduation theses, detailing areas where AI is permitted to facilitate research and where AI is banned.

The university announced in its regulation that Generative AI and other AI-empowered tools are not allowed to be used in choosing research topics, designing research methods, building algorithms and models, analyzing data, presenting results, discussions and conclusions, or summarizing implications in the thesis-writing process.

It also bans AI tools from generating or changing raw data and creating figures based on the results of experiments. The main body, acknowledgments and other sections of the theses cannot be written by AI, nor can AI be used to polish up the language or do translation work.

Moreover, members of thesis defense committees and reviewers are prohibited from using AI tools to review undergraduates' graduation theses.

The document also specified conditions where AI can be leveraged: when the supervisor consents to the use of AI and when the AI-generated content doesn't affect evaluations of students' innovating capabilities. Under these premises, AI tools can be used to help retrieve and review literature, make charts and figures based on already-existing ones and collate references.

The stipulation clearly stated that students should disclose unequivocally in their undergraduate theses what AI tools were used, when and for what purposes, what content and suggestions were generated by such tools, and which parts of their theses were completed using them.

The university said violations of relevant regulations could result in a deduction in graduation thesis scores or disqualification for oral defense. If the improper use of AI constitutes academic misconduct, students will either be unable to apply for the degree, or their already obtained degree will be annulled.

The regulation was introduced as many university students and professors have turned to AI to aid their academic research, while the credibility and academic integrity of such AI-facilitated papers have not yet been fully determined.

In March, a research paper published in a Science Citation Index journal by a professor and colleagues at a university in Beijing was found to contain a commonly used ChatGPT prompter in the article's introduction.

The paper was later retracted due to the "duplication of text and image data" and "concerns that the authors appear to have used a Generative AI source in the writing process without disclosure", which, according to a notice by the journal Surfaces and Interfaces, "is a breach of journal policy".

This year, many universities have announced that they will pilot monitoring the use of Generative AI in undergraduates' graduation theses.

North China Electric Power University has developed an AI detection system to screen the theses of graduate students, the results of which will be provided to the supervisors and the defense committee as a reference, said Zhang Lei, a staff member of the university's graduate school, in an interview with China Youth Daily.

According to Fudan University, the writers are ultimately responsible for the content generated by AI, and AI tools cannot be listed as co-authors of theses.

The regulation aims to stress students' role in ensuring the proper use of AI tools in line with research codes of conduct and preserving the originality and academic integrity emphasized in academic research, Fudan University said.

Xu Nuo contributed to this story.