UK Universities Urged to Return to Traditional Exams to Stop AI Cheating

UK universities must adopt assessment methods that are resistant to artificial intelligence (AI), including traditional paper-based examinations and oral presentations, to prevent cheating and protect academic standards, England’s higher education regulator has warned.

The Office for Students (OfS) said AI posed a “generational risk” to the integrity and reputation of universities. It warned that institutions could face fines and other disciplinary action if they failed to prevent students from misusing the technology.

While recognising the “huge opportunity” AI offers to improve teaching, the regulator said universities needed to do more to prevent cheating from becoming normalised and undermining students’ ability to learn.

Student surveys suggest that more than 90 per cent of students use AI to help complete assignments. In response, some universities have begun redesigning assessments, including reviving oral examinations, known as viva voce, which date back to the medieval era.

In a report published on Wednesday, the OfS said universities' approaches to AI-related academic misconduct “vary widely”. It also acknowledged criticism that the regulator had been slow to establish clear rules since ChatGPT emerged four years ago.

The regulator praised some universities for changing their assessment methods but said too many continued to rely on unsupervised coursework, which remains highly vulnerable to AI misuse.

Students frequently reported that some of their peers “had completed assessments using AI, sometimes in just a few minutes, and received a similar grade to those who had completed the assessment themselves”, the report said.

Many students “see AI as the most effective route to a high grade” and “feel they are at a disadvantage” if they do not use the technology themselves.

Academics also reported regularly seeing assignments written in similar styles and containing the same references. Some responses suggested that “coursework is now overwhelmingly written in part or exclusively with the help of AI”.

Josh Fleming, director of strategy at the OfS, said AI “presents a generational risk to the system” and undermines universities' ability to assess students fairly.

“We need to see far greater efforts to tackle these risks, including far greater use of AI-free assessment.”

Fleming said AI detection software was not sufficiently reliable to be used as the basis for disciplinary action. “It’s much better to design the assessment such that it’s not even an option to cheat,” he added.

The OfS has issued new guidance urging universities to make greater use of secure assessments, including in-person examinations, presentations and oral examinations. Under an overhauled quality regime, universities must demonstrate “that assessments are designed in such a way to be resilient against AI”.

However, the report also highlighted the potential benefits of AI when used appropriately, including providing rapid, personalised feedback to students and helping lecturers develop better teaching materials.

Fleming urged universities to act before regulatory intervention becomes necessary, saying they should “get ahead of this and not wait for the regulator to come and find the issue”.

He praised the University of Bath for adopting a “two-lane approach”, which combines examinations in which students cannot access AI with other assessments explicitly designed on the assumption that the technology will be used.

Universities UK said many institutions “have made substantial changes to how they assess students”.

“This includes more in-person exams, presentations, vivas, process-based assessment and assessments that integrate AI in a controlled way,” it said.

“Different subjects will require different approaches, and universities should have the flexibility to adopt assessment methods that best support learning outcomes.”