England’s Office for Students (OfS) has issued new guidance urging universities to use more AI-free assessments, including in-person exams, presentations and oral examinations. Institutions that fail to assess students’ knowledge and skills reliably could face sanctions. Source Source
That isn’t an order to march every assignment back into an exam hall. Times Higher Education reports that the guidance stops short of mandating AI-free assessment. Where AI is permitted, universities should make the rules clear through accessible policies and transparent academic regulations. Source
Practical tasks and performance-based assessments are also on the regulator’s list. Source
Redesign the assessment, not just the detection
Josh Fleming, the OfS director of strategy and delivery, described AI as a “generational risk” threatening students’ learning and universities’ ability to assess them. “We need to see far greater efforts to tackle these risks, including far greater use of AI-free assessment,” he said. Source
The warning follows Making AI Work for Students, an OfS and Advance HE study involving about 300 students and staff across more than 120 institutions, according to Times Higher Education and the Guardian. Participants described assessments that they believed no longer reliably measured what students knew. Source Source
Students told researchers that peers had completed assessments using AI, sometimes in minutes, and received grades similar to those who completed the work themselves. Those accounts don’t establish a sector-wide cheating rate. Source Source
Nor is the proposed answer simply to buy a better chatbot detector. Fleming told the Financial Times that AI-detection software was not reliable enough to use in disciplinary action. He favoured designing assessments that prevent opportunities to cheat in the first place. Source
The regulatory obligation itself isn’t new. Existing condition B4 already requires effective assessment methods and secure processes; the new advice explains how those expectations apply to AI. Source
There’s still a lane for AI
The FT reports that Fleming praised the University of Bath’s “two-lane approach”: some assessments exclude AI, while others are explicitly designed around its use. The Guardian also reports existing redesigns in which students critique AI-generated output or document how they researched and produced their work. Different tasks, different rules—not one enormous chatbot eviction notice. Source Source
Universities UK told the FT that many institutions had already made substantial assessment changes. “Different subjects will require different approaches,” it said, arguing that universities need flexibility to choose methods that support learning. Source
Speaking in Parliament before the report’s release, King’s College London vice-dean Simon Banks argued that institutions must preserve genuine benefits of remote assessments and build inclusivity into their replacements from the outset. “You can’t retrofit that afterwards,” he said. Source
Comments