"It also presents a generational risk to the system"
England’s higher education regulator has warned that too many university students can use AI to complete assessments without their lecturers finding out.
The Office for Students (OfS) said a significant proportion of university assessments were conducted without supervision, allowing students to use AI tools to produce work to a high standard, sometimes within minutes.
The regulator warned that this posed a “significant risk” to the credibility of higher education qualifications, as institutions could struggle to establish whether students were demonstrating their own knowledge and skills.
Its research, conducted with Advance HE, found that many assessments could be completed using AI with limited involvement from students.
It also concluded that detecting such use was “frequently possible” only with difficulty, as it was not always possible to reliably identify when students had used AI.
The findings were based on the views of 244 higher education staff and 91 students from at least 127 universities.
The OfS identified coursework and remote online examinations as examples of unsupervised assessments where students could access AI tools.
Remote exams became more widely used during the Covid pandemic, with research from Swansea University in 2025 suggesting that 93 UK universities continued to use them in some form.
The regulator said closed-book and practical examinations were among the most effective ways of assessing students without the risk of AI undermining the process.
Josh Fleming, director of strategy and delivery at the OfS, said there were “fantastic” examples of universities making use of the opportunities presented by AI, but argued that there needed to be “far greater use of AI-free assessment”.
He added: “It also presents a generational risk to the system as it stands, threatening to undermine students’ learning and institutions’ ability to assess them.”
The research also highlighted concerns among university staff about how AI was changing student behaviour and the reliability of assessment methods.
The OfS said it had heard of cases where students completed entire assessments within minutes using AI and received similar grades to those who had completed the work without it.
Some lecturers were also receiving large numbers of assignments that appeared almost identical, prompting concerns about whether existing assessment methods remained “fit for purpose”.
The research further identified a “polarised atmosphere” between students who regularly used AI and those who avoided it, including so-called “conscientious objectors” who had environmental and other concerns about the technology.
Kathryn Harrison-Graves, deputy chief executive at Advance HE, said students wanted “greater clarity about what is expected of them”, while staff had “varying levels of confidence and support”.
The report stressed the importance of improving AI literacy among students, both during their studies and after graduating. It also highlighted examples of universities clearly explaining when AI could and could not be used in assessments.
However, it warned that there was “an urgent need” to ensure remote and “non-secure” assessments remained valid despite the availability of AI tools.
It added: “If it becomes possible to complete an assessment using AI without the knowledge and skills being tested, the assessment should be replaced.”
The findings come as universities face growing pressure to adapt their assessment methods, with Universities UK saying institutions were already making changes to address the impact of AI.
A spokesperson for the organisation, which represents more than 140 universities, said institutions were “making significant progress” in adapting to the technology.
The spokesperson said:
“There is further to go but many 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.
“Different subjects will require different approaches, and universities should have the flexibility to adopt assessment methods that best support learning outcomes.”
Lewis Wilson, from the National Union of Students, said AI was also exposing longstanding problems with traditional university teaching and assessment methods.
He argued that essay questions, handouts and slideshows were among the approaches that students and staff “have long said is not fit for purpose”.
