Skip to content
News — 2026

Office for Students calls for in-person exams in bid to reduce AI use in assessments

The OfS has called on universities for an overhaul of assessments to reflect a new AI cheating landscape

By Dmitri Cole, Third year, Law

The Office for Students (OfS) has called for ‘secure’ assessments, including in-person written exams, presentations or vivas, in a bid to secure assessments against AI use - with universities facing sanctions if they fail to do so.  

The persistence of unsupervised assessments after the development of AI poses a ‘significant risk’ to university reputation according to the OfS, as more than 90 per cent of students surveyed claim to use AI to help with assignments.

The regulator further warned that AI had made it ‘close to impossible’ to assess coursework. The OfS highlighted that assessments can be completed to a ‘high standard using AI tools with limited engagement from students,’ sometimes within ‘minutes.’ These students have ‘received a similar grade to those who had completed the assessment themselves.’

One student was quoted as saying: ‘… It is incredibly disconcerting when I plan, and write an essay completely by myself, and get the exact same grade as someone who used AI to write it.’

AI cheating is a problem and no one is asking why
The misuse of Artificial Intelligence in assessments is increasingly common, and so is hustle culture. Megan Pantry explores why these two issues might well be connected.

The OfS further found that there is a widespread belief amongst students that using AI is the ‘most effective route to a high grade,’ and that opting not to use AI would place them at a ‘disadvantage.’

A student was quoted to this effect, saying ‘I’ve had people admit, peer-to-peer, “I wrote this whole essay with ChatGPT.” […] Then they don’t get caught, they get a good grade probably, and then because it worked the first time, they do it on their next assessment.’

There is also a ‘disproportionate weight’ placed on AI detection tools, despite evidence suggesting they don’t work reliably. The OfS described several strategies adopted by students to avoid AI detections, such as removing em-dashes, using text ‘humaniser’ tools, and deliberately misspelling words.

AI for everyone!
By Vilhelmiina Haavisto, Deputy Science and Technology Editor The use of AI is becoming increasingly commonplace in society, but how are we adapting to the major changes it brings? Artificial intelligence (AI) has gone from being a futuristic sci-fi trope to a real-life technology with a myriad of

The report quotes several academics at different institutions describing present assessment methods as ‘unfit for purpose,’ putting ‘academic integrity at serious risk,’ and generating ‘high volumes of submissions that look the same.’

Another staff member claimed that ‘[c]oursework is now overwhelmingly written in part, or exclusively, with the help of AI. Presentation slides fare no better.’

The OfS found that some members of academic staff believe that their institution’s assessments ‘do not accurately measure students’ knowledge and skills.’

The OfS also highlighted the disparity between those paying for better AI models, as ‘a student paying for the most advanced model may receive a higher grade than a student with equivalent knowledge and skills using a cheaper model or not using AI.’ These superior AI tools may evade AI detectors more effectively. Furthermore, better-quality ‘humaniser’ tools can be paid for, which may also make AI detection more difficult.

How AI promotes an unfair academic culture
AI language models are rewriting the rules of academic study for students, posing serious challenges for integrity, fairness and the very purpose of a degree.
Opinion | AI's potential in academia is offset by its risks
Caroline O'Beirne discusses the increase of students' reliance on services such as ChatGPT, and what the role of AI in academia should be.

The OfS report concluded that ‘[i]f it becomes possible to complete an assessment using AI without the knowledge and skills being tested, the assessment should be replaced.’

The OfS report on assessment summarised that ‘effective institutional practice’ in assessments is ‘likely to require’ secure assessments designed to prevent AI use, clear and effective communication to students regarding permitted AI use, and ‘fair and reliable’ academic misconduct proceedings, without excessive weight placed on the output of AI detection tools. 

The OfS has warned universities they may face sanctions for not meeting OfS quality standards. These include fines, bans from charging increased tuition fees, and caps on student numbers.

Featured Image: Epigram / Hanno Sie

Latest