DEV Community

News
News

Posted on

England regulator pushes universities toward AI-free exams as AI use undermines coursework grading

Infographic cover

๐Ÿ“… November 11, 2026 โ€ข The Guardian

England's higher education regulator, the Office for Students (OfS), says AI has made assessing coursework "close to impossible" and urges universities to return to traditional exams and in-person presentations. The position follows an investigation published by The Guardian on November 11, 2026. The regulator also warned that institutions failing to design valid and reliable measures of knowledge could breach conditions of registration and face sanctions, although it stopped short of mandating AI-free assessment outright.

The findings draw on a joint survey by the OfS and Advance HE that gathered views from more than 300 students and staff at 127 institutions. The BBC counts the respondents as 244 staff plus 91 students. Students reported peers completing entire assessments with AI within minutes, often undetected, and receiving grades similar to students who did the work themselves. Staff reported batches of uniform submissions with identical sources and style. The report authors also note that a significant share of assessments remain unsupervised, non-controlled formats where AI is accessible.

OfS strategy and delivery director Josh Fleming called AI "a generational risk to the system", threatening students' learning and institutions' ability to assess, and demanded "far greater use of AI-free assessment". Institutions are already shifting to in-person written exams, oral examinations, process-based assessments and tasks that critique or edit AI output. The move marks a change in regulatory posture, from guidance toward enforcement levers: conditions of registration and sanctions now attach to assessment design itself, not only to teaching quality or finances.

The practical consequence is that assessment design becomes an AI-governance problem. Unsupervised coursework no longer proves knowledge on its own, because AI can produce it within minutes, so the value of a grade now depends on the format of the task rather than only on the student's effort. Universities face a concrete compliance question, namely which formats remain valid when AI is accessible and what evidence of learning they can defend to the regulator.
The spread of vivas, process-based assessment and AI-critique tasks shows institutions testing formats where using AI does not erase the measurement.

The OfS position sets a direction of travel for English higher education, with more supervised, in-person and process-based formats and AI-free assessment as the reference point for reliability. Whether universities can redesign assessment at scale without inflating workload remains open, but the regulator has made clear that the current unsupervised coursework model is no longer considered a valid measure.

Sources: theguardian.com/education/2026/nov/11/ai-close-to-impossible-assess-university-coursework-regulator - bbc.com/news/articles/cw0k0klz2x37o - timeshighereducation.com/news/ofs-pushes-person-exams-protect-sectors-reputation

Key figures

IMPACT checklist


Continue the conversation: https://t.me/AIAutomationArchitecture ยท Bluesky

AI Automation

Top comments (0)