England’s higher education regulator has urged universities to bring back more supervised exams and in-person presentations. The Office for Students (OfS) warned on November 11 that widespread AI use in essays, dissertations and open-book exams could undermine confidence in degrees. Lecturers told researchers that AI had made it “close to impossible” to assess students’ work in some areas.
Universities and colleges could face sanctions if their assessments fail to measure what students know and can do. Under the regulator’s existing Condition B4, providers must assess students effectively and ensure each assessment is valid and reliable. They must also design academic regulations that protect the credibility of their qualifications.
Read More: FBR’s AI will cross-check your tax return. What can it flag?
Josh Fleming, the OfS director of strategy and delivery, said AI “presents a generational risk to the system as it stands, threatening to undermine students’ learning and institutions’ ability to assess them”. He called for wider use of AI-free assessment. He said the speed and scale of change posed a challenge for universities and the regulator. The OfS would continue working with institutions to help them use AI while managing the risks, he added.
Students say classmates used ChatGPT and escaped detection
The OfS and Advance HE gathered evidence from more than 300 students and staff across 127 universities and colleges. The research ran from June to September and examined teaching, assessment and AI literacy.
Students said some classmates used AI to finish assignments within minutes and earned grades similar to peers who completed the work themselves. Some could barely recall what they had submitted. One student recalled classmates admitting, ‘I wrote this whole essay with ChatGPT’. The student said classmates repeated the tactic after escaping detection.
Read More: ChatGPT Locked Chats: PIN Protection and Reset Warning
Lecturers described batches of submissions that cited the same literature and sounded alike. One academic said AI had put academic integrity at serious risk, with students using it to write some or all of their coursework. Another questioned whether existing assessments could still measure students’ knowledge and skills.
Universities turn to oral exams and supervised assessments
Some institutions already use more supervised written exams, live presentations and oral examinations known as vivas. Others ask students to critique or edit AI-generated work. Another approach asks students to document how they researched and produced an assignment, giving lecturers evidence of the work behind the submission.
Universities UK, which represents 147 higher education institutions, said many had already made substantial changes. It cited assessments that allow AI under controlled conditions alongside supervised exams and assignments that track students’ work. The Department for Education said it expected the OfS to enforce existing standards and move quickly to clarify its rules on AI.
