The Office for Students (OfS), England’s higher education regulator, issued guidance on 11 November urging universities to make far greater use of AI-free assessments, including written exams and in-person presentations, the Guardian reported. The guidance follows an OfS investigation into AI use in teaching and assessment. Lecturers told researchers that widespread use of the tools had made some coursework difficult to assess, while the regulator warned that institutions could face sanctions if their methods failed to measure students’ knowledge and skills reliably.
Key points
- The OfS wants greater use of assessments designed to prevent AI use, including exams, presentations, oral examinations and practical tasks.
- Students told researchers that peers had used AI to finish assessments in minutes and received grades similar to those who did the work themselves.
- Under existing condition B4, providers must use effective assessment methods. The regulator could investigate provision where it identifies quality concerns.
- Separate proposals under consultation could restrict fee increases or impose student-number caps on universities that fail to meet quality standards.
OfS guidance favours secure assessment
The OfS says providers should use assessments designed to prevent students from relying on AI where they need to demonstrate their own learning. Its examples include in-person exams, presentations, vivas — oral examinations — performance-based work and practical tasks. The regulator stopped short of requiring every assessment to be AI-free, but said institutions that permit AI use should set out the rules clearly, Times Higher Education reported.
Josh Fleming, the OfS director of strategy and delivery, called AI use in higher education a “generational risk to the system as it stands, threatening to undermine students’ learning and institutions’ ability to assess them”. He called for far greater use of AI-free assessment. The OfS also described approaches that incorporate the technology: asking students to critique or edit AI-generated work, or to explain the steps by which they researched and produced an assignment.
That distinction matters for universities whose courses use coursework to test more than recall. The OfS is asking them to decide where a student must work without AI and where its use can form part of an assessment, with rules that tell students which applies. It says a substantial share of assessments still take place without supervision or controls, in formats that allow students to access AI while completing them.
Advance HE research records minutes-long submissions
The guidance follows a four-month study by the OfS and Advance HE, a charity concerned with university governance and teaching. The Guardian reported that researchers spoke to more than 300 students and staff at 127 universities and colleges. Times Higher Education put the group at some 300 people from more than 120 institutions. Their accounts included students who said peers had completed assessments with AI in minutes yet received grades similar to those earned by students working without it.
The research found that many unsupervised assessments could be completed to a high standard with little engagement from the student, the BBC reported. It also found that AI use in such work was frequently difficult to detect reliably. One member of staff told the researchers that some assessments might be “unfit for purpose” and that pervasive AI use could make assessing students “close to impossible in some areas”. Other staff described receiving submissions with similar writing and references.
The study also recorded disagreement over AI’s place in teaching. Some students objected to its use in marking and valued feedback from teaching staff. Some staff said AI-generated feedback could provide a starting point that left them more time to speak with students. Kathryn Harrison-Graves, Advance HE’s deputy chief executive, said students wanted clearer expectations, while staff had varying levels of confidence and support.
Universities UK, which represents more than 140 institutions, said many universities had changed their assessments, using more in-person exams, presentations and vivas alongside assignments that incorporate AI under controls. It argued that subjects need different approaches and that universities should retain flexibility to choose methods suited to what students are meant to learn. Lewis Wilson of the National Union of Students called for assessment practices to be developed with students rather than for older formats to be retained simply because they are traditional.
Condition B4 gives the OfS enforcement powers
Condition B4 of the OfS’s existing quality requirements calls for effective methods of assessing student learning. The regulator says concerns identified through providers’ submissions or other information could result in a lower quality award and inform a decision to investigate a provider’s courses further. Institutions that fail to design valid and reliable ways to test knowledge and skills could breach their conditions of registration.
The Daily Mail described the guidance as threatening fines for universities that allow AI to compromise assessments. The BBC reported a separate, proposed set of consequences now under consultation: universities that fail to meet OfS quality standards could be barred from raising tuition fees or have limits placed on student numbers. Those proposed restrictions are distinct from the quality requirements already in force.
A Department for Education spokesperson said universities were already responsible for preventing academic misconduct, including AI misuse. The department expects the OfS to enforce existing standards and to make its regulatory framework for AI explicit and durable.
