England’s university regulator urges more in-person exams as AI threatens assessment
The Office for Students says widespread AI use is putting degree credibility at risk and warns institutions that inadequate assessment methods could lead to sanctions.
By Hiroshi Ferrari ·
England’s higher education regulator has urged universities to make greater use of traditional exams and in-person presentations amid concerns that artificial intelligence is undermining their ability to assess students. An investigation found that staff questioned whether some coursework could still reliably demonstrate what students knew and could do.
The Office for Students said widespread AI use in essays, dissertations and open-book examinations threatened the credibility of degrees awarded by England’s universities. It warned universities and other higher education providers that assessment methods which failed to provide valid, reliable measures of knowledge and skills could breach registration conditions and result in sanctions.
The findings came from research conducted by the regulator and Advance HE, a charity focused on university governance and teaching. Researchers interviewed more than 300 students and staff across 127 universities and colleges during 2026, examining institutional AI policies alongside the technology’s use in teaching and assessment.
Students described peers completing assessed work with AI, sometimes within minutes, and obtaining grades similar to those awarded to students who had worked without it. Some said classmates could not remember basic details of assignments they had submitted. Accounts also included students boasting that they had used AI without being detected, with successful attempts encouraging them to repeat the practice in later assessments.
Staff reported receiving numerous submissions with similar writing styles and references to the same literature. One academic said coursework was now widely produced either partly or entirely with AI assistance. Others questioned whether their institutions’ assessments remained suitable or accurately reflected students’ abilities, with one staff member describing assessment in some areas as becoming almost impossible.
Josh Fleming, the OfS director of strategy and delivery, said AI posed a “generational risk” to higher education by threatening both learning and institutions’ capacity to assess it. He called for substantially stronger action, including wider use of assessments completed without AI, and said the regulator would continue working with universities and colleges to help them benefit from the technology while managing its risks.
Many institutions have already changed their assessment practices, according to the regulator. Approaches include more written examinations conducted in person and oral examinations, known as vivas. Other methods ask students to critique or edit AI-generated material, or explain the research and production process behind their submitted work.
The Department for Education said universities were expected to treat the issue seriously. A spokesperson also called on the OfS to enforce existing standards effectively and move quickly to ensure its regulatory framework addressed AI explicitly, with strong and durable requirements.
