Ph.D. Student Sues University of Minnesota Over AI Expulsion

MINNEAPOLIS, Minn. — Haishan Yang, a third-year Ph.D. student, has filed a lawsuit against the University of Minnesota and one of its professors after he was expelled for allegedly using generative artificial intelligence (AI) on an exam.

Yang’s expulsion in January marks a significant instance as he appears to be the first student terminated for suspected academic dishonesty related to AI. The lawsuit, filed in February, contends that Yang was wrongfully accused of violating the university’s scholastic dishonesty rules during his preliminary exam taken in August.

“I didn’t use AI in my exam, and the accusations are baseless,” Yang said. “I hope to clear my name and continue my studies.”

In addition to the lawsuit, Yang also filed a complaint with the Minnesota Department of Human Rights, claiming that he faced discrimination based on his national origin. Originally from Quanzhou, China, Yang came to the University to pursue a doctorate in health economics after earning a prior doctorate in economics from Utah State University.

Yang stated that he initially utilized AI tools, including ChatGPT, to check his grammar as a non-native English speaker but chose against using them for his exam. Documents shared with the Minnesota Daily indicate concerns raised by four faculty graders, who believed that a certain section of Yang’s exam showed signs of potential AI assistance. They noted that an acronym not introduced in preparatory materials appeared in his responses.

Professor Peter Huckfeldt expressed in a hearing letter, “I was struck by the similarities between the two that seemed extremely unlikely to be coincidental.”

Yang alleges that professor Hannah Neprash, who was involved in the review, altered parts of a ChatGPT-generated response, such as removing a summary paragraph and changing text formatting. He pointed out that Neprash failed to present a verifiable link to the original AI response, only submitting a PDF of her findings.

Neprash declined to comment on the matter due to student privacy laws. A February 21 court filing argued that the modifications identified by Yang were “trivial differences, even if intentionally made.”

During the hearing, Roxanne Krietzman, assistant director of student advocacy for the University’s Student Advocate Services, noted that the questioned section constituted only 6.6% of Yang’s entire exam and that AI detection tools indicated a lower likelihood of AI use on other questions.

Yang also contended that the methodology used for AI detection may not be reliable for evaluating non-native English speakers’ work. Louie Giray, an assistant professor of communication at Mapúa University in the Philippines, explained that detection systems assess writing patterns in comparison to typical English usage, often flagging straightforward language employed by non-native speakers.

Furthermore, a study from the University of Washington revealed that AI detection software, like Turnitin, produced a 25% false positive rate when evaluating work that did not employ AI. Giray believes there will not be a fool-proof AI detector in the future as AI technology continues to advance rapidly.

The University of Minnesota has refrained from commenting on the lawsuit due to federal and state privacy laws. University spokesperson Jake Ricker declined interviews on behalf of the faculty members implicated in the lawsuit.

According to the University’s definition of scholastic dishonesty, “the unauthorized use of online learning support and testing platforms” is prohibited. If Yang’s lawsuit is resolved in his favor, he hopes to complete his program but indicated he would pursue a different academic track.

Currently on travels in Africa, Yang mentioned he is actively seeking faculty opportunities in economics globally. “I’m still going to try to survive, try to find a new opportunity,” he said.