England's higher education regulator has warned that a substantial portion of university students can deploy artificial intelligence to finish assessments without detection, raising serious questions about the integrity of degree qualifications. According to BBC News, the Office for Students (OfS) found that many unsupervised assessments can be completed to a high standard using AI in a matter of minutes, posing what it describes as a "significant risk" to institutional credibility.
The OfS, working with Advance HE—a charity that recognises teaching excellence—conducted research involving 244 higher education staff members and 91 students across at least 127 universities. The investigation revealed that the OfS's joint research also drew on roundtables, interviews and an online survey, with input from senior institutional leaders and student representatives, providing a comprehensive picture of how AI is reshaping assessment practices.
How widespread is unsupervised assessment?
A significant proportion of assessments operate in unsupervised, non-controlled formats where students can freely access AI tools while completing them. These unsupervised assessments encompass coursework and remote online examinations, which became widespread during the Covid-19 pandemic. Research from Swansea University indicates that 93 universities across the UK continue to employ remote online exams in some capacity, demonstrating how embedded this practice has become in higher education.
The OfS research found that many of these assessments can be completed to a high standard using AI tools with minimal genuine engagement from students. Regulators also concluded it is frequently possible for students to use AI in ways that cannot be reliably detected by staff, creating a fundamental challenge to assessment validity.
What specific problems have been identified?
The regulator has documented cases where students completed entire assessments within minutes using AI and received comparable grades to those who had not used such tools. Simultaneously, some staff members reported receiving numerous assignments that appeared nearly identical, raising concerns about whether their institution's assessments were genuinely fit for purpose.
A polarised atmosphere has emerged within some institutions, with staff describing tension between students who openly use AI and what they termed "conscientious objectors" who avoid the technology on environmental and other ethical grounds. This division reflects broader uncertainty about acceptable use of AI in academic work.
What assessment methods are most effective?
The OfS identified closed-book examinations and practical assessments as the most reliable forms of evaluation, as these formats inherently limit opportunities for undetected AI use. Josh Fleming, director of strategy and delivery at the OfS, acknowledged that some universities have seized opportunities presented by AI but called for "far greater use of AI-free assessment."
"It also presents a generational risk to the system as it stands, threatening to undermine students' learning and institutions' ability to assess them,"
Fleming stated. Universities UK, representing more than 140 institutions, countered that many have already "made substantial changes" to assessment methods. A spokesperson outlined the adaptations underway: "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."
What support do staff and students need?
Kathryn Harrison-Graves, deputy chief executive at Advance HE, highlighted a critical gap in institutional readiness. Students want "greater clarity about what is expected of them," she said, while staff possess "varying levels of confidence and support" in managing AI-related assessment challenges. The research emphasised that AI literacy for students is essential both during and after university, with some institutions already communicating clearly to students where AI can and cannot be used in assessments.
The OfS stressed there is "an urgent need" to ensure that remote or "non-secure" assessments "maintain their validity despite the availability of AI tools, meaning they effectively assess the knowledge and skills they are designed to measure." The regulator added a stark principle: "If it becomes possible to complete an assessment using AI without the knowledge and skills being tested, the assessment should be replaced."
How does this connect to broader concerns about AI in education?
This challenge is not unique to England. Denmark has tightened rules on AI cheating in upper secondary schools, requiring students to orally defend essays and implementing stricter exam monitoring. Similarly, concerns about false accusations have emerged; a St Andrews medical student was wrongly accused of using AI in his dissertation, triggering hearings that delayed his exams before the finding was overturned. These cases underscore the complexity of detecting and responding to AI use in academic work.
Educational institutions are also investing in AI literacy programmes. Scotland's first dedicated AI literacy teacher warns pupils to question and verify AI outputs rather than trust them blindly, teaching students to spot deepfakes and recognise hallucinations. This approach reflects a shift toward equipping learners with critical thinking skills rather than simply restricting technology access.
What regulatory action is planned?
The OfS has proposed new enforcement mechanisms to drive institutional change. Universities that fail to meet OfS quality standards could soon face restrictions on charging increased tuition fees or have caps imposed on their student numbers. These proposals are currently under consultation, signalling the regulator's determination to ensure assessment integrity across the sector.
The research builds on OfS work dating to early 2025, including earlier staff and student discussions about AI's effects on higher education, indicating this is an evolving priority for the regulator.
What happens next?
The OfS and Advance HE have scheduled an online discussion of the research findings for 16 November 2026; a report will be published shortly to inform the discussion. This event will bring together sector leaders to discuss implementation of recommendations and next steps for managing AI in higher education assessment.
Lewis Wilson, from the National Union of Students, offered a contrasting perspective, arguing that AI is exposing the limitations of "traditional" assessment methods that students and staff "have long said is not fit for purpose"—including essay questions, handouts and slideshows. He called for institutions to "get rid of the outdated methods, that have been kept around in the name of tradition, and instead encourage the innovative practice we see across the sector, co-designed with students based on what they want from higher education."
Key Facts
- The OfS research involved 244 higher education staff and 91 students from at least 127 universities across England.
- Documented cases show students completing entire assessments in minutes using AI and receiving grades comparable to those who did not use the technology.
- Closed-book and practical exams are identified as the most effective assessment formats for preventing undetected AI use.
- Universities UK reports that many institutions have already implemented substantial changes, including more in-person exams, presentations, vivas, and controlled AI integration.
- The OfS is consulting on proposals that could restrict tuition fee increases or impose student number caps on universities failing to meet quality standards.
