Universities “At Risk of Overassessing” in Response to AI

The number of assessments set by universities is steadily rising, but there are worries this could result in student burnout and prove counteractive if implemented without centering learning.

A recent report by the U.K.-based Higher Education Policy Institute (Hepi) and Advance HE found that assessments at U.K. institutions have risen to 5.8 summative assignments and 4.1 formative assignments per semester in 2025, compared to five summative assessments and 2.5 formative assessments in 2020.

Josh Freeman, policy manager at Hepi and co-author of the report, said the advent of AI is “reducing the accuracy of assessments as a measure of students’ performance,” prompting universities to re-evaluate their examination methods.

“It’s possible that course organizers are assessing students more to improve the confidence they have in their assessments,” he said.

“It’s also possible that, as they redo assessment models, which may have remained the same for a long time, they are switching to alternative models of assessment—for example, those that assess students on an ongoing basis, rather than simply once at the end of the year.”

However, rising numbers of exams risks universities “overassessing” students, he added, as “students now face an intense battle over their time,” noting that the number of hours that students spend studying has fallen.

“[Many are making] sacrifices around social activities, sports and societies. These ‘extra’ activities are the first to go when students are squeezed and would probably be cut further if the academic elements of university become more demanding.”

Some 68 percent of students in the U.K. are now undertaking part-time work during term time, a record high, largely in response to cost-of-living pressures.

Michael Draper, a professor in legal education at Swansea University and chair of the university’s academic regulations and student cases board, said that some universities have begun supplementing assessments with “some form of in-person assessment” to counteract AI “credibility concerns.” But “that of course does lead to perhaps overassessment or more assessments than were in place before.”

“Students have got so many competing claims on their time, not just in relation to work, but care responsibilities and work responsibilities, that you run the risk of student burnout,” he continued.

“That is not a position you actually want to be in. You want to make sure that students have got a fair opportunity to work consistently and get the best grade possible. You want students to have a chance to reflect upon their feedback and then to demonstrate that in other assessments, but if they’re being continuously assessed, it’s very difficult to have that reflection time.”

However, Thomas Lancaster, principal teaching fellow in the Department of Computing at Imperial College London, speculated that a rise in the number of exams could be a sign that assessments are being “split into smaller stages,” with more continuous feedback throughout the process, which could also simultaneously have benefits for counteracting AI use.

“This is something I’ve long recommended in response to contract cheating, where it’s good practice to see the process, not just the final product. So I do hope that the revised assessment schedules are being put in place to benefit the students, rather than purely as a response to AI.”

While breaking assessments down could prove beneficial to student learning, Drew Whitworth, reader at the Manchester Institute of Education, questioned, “How does one count what constitutes ‘separate’ assessments?”

“If a grade is given partway through this process … this is actually quite helpful for students, answering the question ‘How am I doing?’ and giving them a pragmatic reason to show [their work and that they are working] in the first place.

“But does this count as a separate assessment or just part of a dialogue taking place that helps students develop better work in response to a single assessment?”