← All guides

Mobile Data

Regulators say that artificial intelligence has made the assessment of university coursework “virtually impossible.”

The Office for Students in England has stated that higher education institutions should return to traditional examinations and in-person presentations.

An investigation by the UK’s higher education regulator suggests that universities should return to traditional examinations and in-person presentations for student assessment, following complaints from lecturers that artificial intelligence has made the evaluation of coursework “virtually impossible.”

The Office for Students (OfS) has stated that the credibility of UK university degrees is under threat due to the widespread use of AI in essays, dissertations, and “open-book” examinations. Some students have reported that their peers boast about using AI without being detected by faculty members.

The regulator has warned universities and other higher education institutions that failing to devise “effective and reliable measures to assess student knowledge and skills” in response to the rise of AI could constitute a breach of their registration conditions, potentially leading to sanctions.

Josh Fleming, Head of Strategy and Delivery at the Office for Students (OfS) in the UK, stated that the use of artificial intelligence in higher education “poses a generational risk to the existing higher education system, with the potential to undermine both student learning and the ability of institutions to assess students.”

Fleming added: “We need to see greater efforts to address these risks, including an increased use of assessments that do not rely on artificial intelligence.”

“The scale and pace of both the challenges and the opportunities present unprecedented issues for the sector, as well as for us as a regulator. We will continue to work with universities and colleges to help them harness the benefits of artificial intelligence while mitigating the risks.”

The Office for Students (OfS) in the UK states that many institutions have redesigned their student assessment methods to reduce the potential for the misuse of artificial intelligence. This includes a greater reliance on in-person written or oral examinations (vivas). Other approaches involve asking students to critique or edit AI-generated content, or implementing “process-based” assessments that require students to detail how they conducted their research and completed their assignments.

A spokesperson for the UK Department for Education said: “We expect universities to take this issue seriously, and we look to the Office for Students to effectively enforce existing standards and act swiftly to ensure that its regulatory framework regarding the use of AI is clear, robust, and enduring.”

A spokesperson for Universities UK, which represents 147 higher education institutions, remarked: “While there is still a long way to go, many institutions have already made substantial reforms to how they assess students. This includes an increase in in-person exams, presentations, oral assessments, process-based evaluations, and assessments that incorporate AI in a controlled manner.”

This report—jointly published by the Office for Students (OfS) and Advance HE, a charity dedicated to supporting governance and teaching in higher education—draws on interviews with over 300 students and staff from 127 universities and colleges, conducted this year to evaluate policies and practices surrounding the use of AI in teaching and assessment.

One academic told the researchers: “AI poses a serious threat to academic integrity… the vast majority of coursework is now completed, either in part or entirely, with the help of artificial intelligence.”

Another university staff member remarked, “I think we are coming to realize that some of our assessment methods may no longer be appropriate, and that the widespread use of artificial intelligence could make it nearly impossible to assess students effectively in certain areas.”

The survey revealed that the use of artificial intelligence is highly prevalent. Students reported that their peers often use AI to complete assignments—sometimes in just a matter of minutes—and frequently receive grades comparable to those of students who completed the work without such tools.

The report notes that “some students have observed that their peers are unable to recall even the basic details of the assignments they have just submitted.”

One student told researchers: “I’ve heard someone openly admit, ‘I used ChatGPT to write the entire essay’—and they weren’t caught; they likely even got a high grade. Because it worked the first time, they just did the same thing again for the next assignment.”

Academics have also reported receiving a large number of papers that share identical formatting, cite the same sources, and exhibit a consistent writing style. Others have expressed the view that the assessment methods currently employed by their institutions are either unsuitable or incapable of accurately measuring students’ knowledge and skills.