When a professor at Brown University allowed his students to take their March midterm exams at home, he was alarmed when 40 of the test results came back with perfect scores.
It was the first time in more than three decades of teaching at the Ivy League university that Roberto Serrano had abandoned in-person testing. Mindful of the recent trauma of the 13 December shooting that led to the killing of two students and injured nine others, he had made an exception.
But the test results, which had an average score of 96 per cent – roughly 25 per cent higher than the average grade – led Professor Serrano to suspect that most of his class had used AI to cheat.
‘Certain assignments that once required sustained effort over hours or even weeks can now be completed almost instantly’ (Getty/iStock)
In an effort to give them the benefit of the doubt, he decided to hold the final exams in person. The average score was just 48.6 per cent, with only one student scoring higher than on the midterm.
An investigation into the incident by the Brown University Standing Committee is currently underway, but Professor Serrano has already pledged to never allow at-home testing again.
Princeton University also recently voted to end a 133-year-old tradition of unsupervised exams, while three-quarters of respondents to a survey from the American Association of Colleges and Universities said they were concerned about students using generative AI tools like ChatGPT to cheat.
A separate report on trust in higher education by Yale University in April described AI as a “truly unprecedented challenge” for higher education institutions.
“Certain assignments that once required sustained effort over hours or even weeks can now be completed almost instantly,” the report noted. “Faculty across the university are scrambling to redesign syllabi and assessments. Whatever its promise, AI in its current use on campus undermines the expectations of focused, disciplined thinking that have long been the standard features of a rigorous education.
“Moreover, it is clear that the rapid technological changes of our moment are contributing to declining public trust in the very idea of human expertise.”
The report also acknowledged that it must balance its approach to AI in classrooms with the need to prepare students for future careers in which they will be required to use the technology.
Earlier this month, the University of Chicago Law School announced a ban on laptops and smartphones for incoming students in an effort to combat AI usage – though it also unveiled a plan to teach how to use AI effectively. The school’s new AI strategy even includes the words: “Technology is changing too fast.”
The new policy will prevent incoming students for the 2026-2027 academic year from accessing any electronic devices during classes, while all examinations will be in-class without access to the internet, electronic files or apps.
By reverting to pen and paper, the law school hopes to encourage the “essential human skills” required to be a lawyer.
“We need to ensure that our students actually learn to think critically, strategically, and independently without relying on AI; but we also must face the reality that AI tools are already widely available to our students, and our graduates will be expected to be prepared to use them in legal practice,” the law school’s new AI strategy states.
The objective is not to make assessment AI-proof. That is probably impossible. The objective is to make learning sufficiently visible that authorship is not reduced to guesswork
Dr John Milliken, lecturer in education, University of Ulster
One of the biggest challenges is ensuring there has been original human effort on core components like coursework, which typically require hours or days of effort that can not be done in a supervised classroom setting.
One potential solution is AI detection tools that can identify writing that has been generated by artificial intelligence tools like ChatGPT. However, these remain fundamentally flawed, with the development of AI models often outpacing the apps designed to catch them.
A report last week by the Higher Education Policy Institute (HEPI) found that such tools are often inaccurate and unreliable. One of the most rampant causes of false positives is proving to be writing by non-native English speakers, which is incorrectly flagged as AI.
AI detection tools are seen as a possible solution, but many are struggling to keep up with the pace of the technology, or producing false positives (AP)
Several leading universities, including UCLA in the US and the University of Waterloo in Canada, have already abandoned popular AI detection tools due to the risk of wrongful accusations.
“The alternative to detection is not the abandonment of academic integrity but the pursuit of it through approaches that work,” the 20 July report from HEPI stated, calling for a radical redesign of programmes so that they no longer relied on essay-based submissions. Instead, institutions should aim for “process-based assessment, staged submissions, oral defences and AI literacy embedded in curricula.”
Rather than rely on AI detection tools, academics are increasingly calling for the revival of assessment methods that have gradually disappeared over the past two decades, returning to handwritten examinations, oral assessments and supervised classroom testing.
“The proper response is not to build more elaborate systems for inspecting finished products. It is to redesign assessments so that learning becomes visible throughout the process,” says Dr John Milliken, a lecturer in education who spent 30 years teaching at the University of Ulster.
“The objective is not to make assessment AI-proof. That is probably impossible. The objective is to make learning sufficiently visible that authorship is not reduced to guesswork.”
The recent scandal at Brown University may well serve as a reckoning for universities. The largest known AI-assisted cheating incident serves as a demonstration of what could happen if a new approach to teaching is not adopted.
“I believe the arrival of AI has been like a tsunami for all of us,” the prize-winning economist told Fortune shortly after the test results were made public.
“It’s caught everybody unprepared. But in my humble opinion, silence is the worst treatment for this problem.”