AI has made it ‘close to impossible’ to assess university coursework, regulator says
Reported by Daily Mail, The Guardian, The Decoder and 1 more outlets4 sources ↓
England’s Office for Students says universities should use more in-person exams and presentations because AI can make coursework grades unreliable. It warns that failing to assess students’ knowledge and skills effectively could breach registration conditions and lead to sanctions.
Key points
- The OfS and Advance HE report draws on more than 300 students and staff at 127 universities and colleges
- Students said peers used AI to complete assessments quickly and received grades similar to students who did the work themselves
- Some students reportedly could not recall basic details of work they had submitted
- Universities are also using vivas, process-based assessments and tasks that ask students to critique or edit AI output
Sources
- Daily MailRead ↗AI is making degree assessments 'unfit for purpose' with students completing them 'in minutes' for top grades, lecturers say - as watchdog calls for more traditional exams
- The GuardianRead ↗AI has made it ‘close to impossible’ to assess university coursework, regulator says
- The DecoderRead ↗UK regulator threatens universities with sanctions unless they adopt AI-proof exam formats
- Tech TimesRead ↗AI-Free Exams: English Regulator Warns AI Coursework Can Earn High Marks and Escape Detection
Details
The OfS says AI use in essays, dissertations and open-book exams threatens the credibility of degrees. It warns that institutions could face sanctions if they fail to use valid and reliable ways to assess students’ knowledge and skills.
The joint OfS and Advance HE report is based on interviews with more than 300 students and staff at 127 universities and colleges. Students reported that peers sometimes completed assessments with AI in minutes and received grades similar to those of students who worked without it; some reportedly could not recall basic details of submitted work.
The OfS says many institutions have already changed assessment methods. Reported approaches include in-person written and oral exams, tasks requiring students to critique or edit AI-generated material, and assessments that document how students researched and produced their work.
Written from coverage as of
Images
- The Guardian
- The Decoder
- Tech Times