Brown University economics professor Roberto Serrano criticized the university’s delayed action in addressing a significant AI cheating scandal earlier this year. In an opinion piece for The Free Press, Serrano detailed his decision to change his in-class exams to an online, take-home format after a campus shooting.
This change resulted in a surge of enrollment from 30 to 86 students. The online midterm saw an average grade spike to 96 from previous averages of 65 to 80. Forty students achieved perfect scores, with answers resembling those provided by AI tool ChatGPT.
Brown University professor Roberto Serrano reported widespread AI usage among his students to earn midterm scores this year.
Reacting to this, Serrano reverted the final exam to an in-person format, with conditions to nullify midterm scores if inconsistent with final exam results. Only 59 students attended the final exam, with an average score of 48.6, and a top score of 95. Twenty-seven students, including 22 who aced the midterm, dropped the course before the final.
Serrano communicated his findings to Brown’s Academic Code Committee and The Brown Daily Herald but received no acknowledgment until the issue gained public attention. He emphasized the necessity of upholding academic integrity in the AI era, warning that ignoring cheating fosters it.
Brown University responded that they had been responsive to Serrano’s concerns, detailing steps taken to formally address the allegations. According to the university, formal proceedings began after Serrano provided the necessary details in July.
Serrano believed the university acted only after his story became public, noting challenges that educational institutions face in integrating AI constructively. He stated that students misusing AI in his exam were not alleviating pressure but intentionally cheating for perfect grades.
Serrano dismissed claims that competition justifies cheating, asserting that his exam’s generous timeframe negated such pressures. He argued that AI’s ease of facilitating cheating necessitates stronger incentives for ethical behavior.