DistantNews
Support us
Brown University Probes AI Cheating Allegations After Exam Scores Plummet
๐Ÿ‡ฐ๐Ÿ‡ท South Korea /Culture & Society

Brown University Probes AI Cheating Allegations After Exam Scores Plummet

From Dong-A Ilbo · () Korean

Translated from Korean, summarized and contextualized by DistantNews.

At a glance

News Sources not specified Under investigation
  • US Ivy League university Brown is investigating allegations of widespread cheating using artificial intelligence.
  • Students in an economics class who averaged 96% on an online midterm scored an average of 46% on the in-person final exam.
  • A professor submitted evidence of AI-generated answers similar to student submissions, prompting the university to launch an investigation into academic dishonesty.

Students at the prestigious Brown University are under scrutiny for allegedly using artificial intelligence to cheat on exams. In an economics course, students who achieved an average of 96% on a remote midterm exam saw their scores plummet to an average of 46% when the final exam was administered in person.

The average score was usually between 65-80%, but this exam was actually more difficult than previous ones.

โ€” Roberto SerranoProfessor Roberto Serrano described the performance difference in his economics class.

Roberto Serrano, a professor of economics at Brown, submitted evidence to the university's academic ethics committee showing a dramatic drop in student performance. He noted that the midterm, conducted remotely due to a campus shooting, saw most students score highly. However, upon inputting exam questions into ChatGPT, Serrano discovered that the AI's generated answers closely resembled those of his students, though often lacking proper reasoning.

When the final exam was shifted to an in-person format, 18 students withdrew from the course, and 9 did not take the exam. Of the 59 students who completed the final, 37 saw their scores drop by more than 40 points compared to the midterm. Two students who had scored 100% on the midterm received grades of 19% and 16.5% on the final.

The answers are correct, but the reasoning process is quite flawed and verbose.

โ€” Roberto SerranoProfessor Serrano analyzed the AI-generated answers compared to student submissions.

Serrano believes the situation highlights the ease with which students can be tempted to use AI, given the near-zero cost of cheating. He praised students who maintained or improved their scores through their own efforts, calling one student who improved from 55% to 59% "respectable" for their integrity. This sentiment was echoed by Glass AI founder Tom Hencke, who suggested hiring the student, a notion Serrano agreed with.

The cost of cheating has become virtually zero, making it easy for students to be tempted.

โ€” Roberto SerranoProfessor Serrano commented on the accessibility of AI tools for academic dishonesty.

Brown University has stated it takes all allegations of academic dishonesty very seriously and will proceed with an investigation. Serrano plans to discontinue remote exams and is considering removing AI-assisted assignments from grade calculations.

Hire this person immediately.

โ€” Tom HenckeGlass AI founder Tom Hencke commented on a student who improved their grade slightly without apparent AI use.
DistantNews Editorial

Originally published by Dong-A Ilbo in Korean. Translated, summarized, and contextualized by our editorial team with added local perspective. Read our editorial standards.