The Haishan Yang case: Unraveling the Complexities of AI in Academic Integrity

In an era where artificial intelligence is reshaping educational landscapes, the case of haishan Yang presents a cautionary tale that might redefine academic integrity protocols across universities worldwide. How can institutions balance fairness and innovation in an age where AI tools are ubiquitous?

Haishan yang,a graduate student at the University of Minnesota pursuing a PhD in Health Services Research,Policy,and Management,found himself at the center of a high-profile controversy. Believed to be the first student expelled from the University of Minnesota for allegedly using AI to cheat, Yang is now suing the institution, claiming his innocence. The controversy began in August 2024, when Yang, working remotely from Morocco, completed a crucial preliminary exam. while permitted to use notes and course materials, AI was explicitly prohibited. I think I did perfect, he stated in an interview. I was very happy.

The University’s Case

Weeks later, Yang received an email informing him that he had failed. The grading professors accused him of using an AI program, such as ChatGPT, to generate his answers. This accusation led to a student conduct review hearing. The university presented evidence based on the professors’ assessment of Yang’s exam responses. All four professors expressed significant concerns that the writing style didn’t match Yang’s previous work and included concepts not covered in class. Professor Hannah Neprash directly compared Yang’s answers to those generated by chatgpt, finding similarities in structure and language. Specific headlines and bullet points were cited as particularly suspicious.

Professor Peter Huckfeldt noted in a letter to the hearing commitee, “I was struck by the similarities between the two that seemed extremely unlikely to be coincidental.” The use of the acronym PCO (primary care organization) in Yang’s answers further fueled suspicion. While Yang argued the acronym is common in established journals, the professors testified they had never encountered it before. Significantly, the acronym also appeared in ChatGPT’s responses to the same exam questions.

Adding to the university’s case was a separate allegation from a year prior, involving a homework assignment. Professor Susan Mason wrote in a letter to the Office of Student Conduct that one of Yang’s paragraphs concluded with a “note to self” having mentioned that, “re wriet it (sic), make it more casual, like a foreign student write but no ai.” This comment,the university argued,suggested an awareness of AI use and an attempt to mask it.

Yang’s Defense and the Lawsuit

Yang vehemently denies all accusations of AI use. The details of his defense and the specifics of his lawsuit against the University of Minnesota remain undisclosed at this time. However, the case raises significant questions about the reliability of AI detection tools, the fairness of academic disciplinary processes, and the broader implications for students and institutions in the age of readily available AI writing assistants.

A separate incident involving an economics exam further complex the situation. professor Mason, in a witness statement, noted, “It seemed clear . . . that Haishan had used AI to generate this answer and had accidentally left either a note to himself or a prompt for AI in his answer.” Despite the initial allegation, the professor ultimately dropped the accusation, and yang received a university warning. Though, the university’s examination didn’t end there.Yang’s answers were later run through GPTZero, an AI detection tool known for its inconsistencies. this added layer of scrutiny ultimately led to a more formal hearing.

“I did not use ChatGPT on the test,” Yang stated repeatedly in interviews. He contends that certain professors targeted him, alleging in his lawsuit that one professor edited chatgpt answers to make them resemble his own work. When asked if he believed this was a conspiracy, Yang responded, “My advisor, Brian Dowd, certainly believes so.”

Professor Dowd, while declining an interview, wrote a letter to the university committee describing Yang as “the best read student he’d ever encountered.” He deemed the evidence against Yang “inconclusive” and noted, “In over four decades in our Division, I never have seen this level of animosity directed at a student. I have no description for that animosity.”

The conflict appears rooted in a prior dispute between Yang and certain faculty members. The university had previously cut Yang’s financial support following claims of poor performance and inappropriate behavior as a research assistant.“The graduate director told me,in person,that I should consider to quit,” Yang recalled. He appealed the decision with support from Professor Dowd,who described the department’s treatment of Yang as “an embarrassment.” The university later apologized and reinstated his funding,contingent on him not pursuing legal action.

yang believes this history influenced the cheating allegations. He offers two explanations for the similarity between his answers and those generated by ChatGPT. first, he suggests that ChatGPT’s training data likely includes the same textbooks and reports he used in his studies. “It means my answer, probably similar to hundreds or thousands, millions of papers ChatGPT has absorbed,” he explained. Second, he maintains that a professor altered the ChatGPT-generated answers.

yang discovered discrepancies between the ChatGPT answers initially shared among professors and those presented at the hearing. He highlighted these differences, stating, “This is the ethical question: They can keep generating and generating and generating and in some version, ‘Wow, it’s more similar.'” He claims to have identified ten such discrepancies.

The five-member panel, after reviewing all evidence, unanimously concluded that Yang more likely than not cheated. This resulted in his expulsion and the cancellation of his student visa. in January,Yang filed state and federal lawsuits against Professor Hannah Neprash and other university personnel,alleging altered evidence and a lack of due process. He admits to using ChatGPT to assist in drafting these lawsuits.

The defendants have yet to formally respond to the lawsuits.Multiple professors involved declined to comment,and a university spokesperson,Jake Ricker,Senior Public Relations Director,stated,“Federal and state privacy laws prevent the University of Minnesota from public comment on individual student disciplinary actions. As in all student discipline cases, the University carefully followed its policies and procedures, and actions taken in this matter were appropriate. The best source for the University of Minnesota’s outlook on this matter will be in our court filings.”

The outcome of this case will undoubtedly have far-reaching consequences for higher education, setting a precedent for how universities address the challenges and opportunities presented by artificial intelligence in academic settings. The legal battle promises to be closely watched by students,faculty,and administrators across the country.