Durham Professor Quits to Expose AI Essay Cheating Gap

Andy Hamilton, Professor of Philosophy at Durham University and Chair of the Board of Examiners for Philosophy since 2016, resigned that post to publicly call out what he says is an active institutional failure: students using professional-grade AI tools like ChatGPT or Claude are earning first-class results while honest students are penalised by comparison, and Durham's leadership is doing nothing meaningful about it. The structural problem is the COVID-era switch from supervised sit-down exams to at-home computer assessments. That shift removed the conditions under which academic integrity could be enforced. Turnitin and similar plagiarism detection tools cannot reliably identify AI-generated essays, and Hamilton notes that AI companies are actively closing the gap: the detectable "tells" — such as hallucinated references — are being "aggressively stamped out" by those same companies, making detection harder over time. Essentially: (Durham University, Turnitin) are on opposite sides of a widening gap the institutions are not closing. - Affluent students with access to paid, professional AI versions gain a structural advantage over peers using free tools or none at all. - Hamilton argues it is now impossible to mark essays fairly under the at-home system in the era of ChatGPT. - His proposed fix is straightforward: return to sit-down examinations as the primary assessment method, with vivas and suspension of anonymous marking as interim measures. The deeper issue is not detection technology but assessment architecture — universities adopted a model during COVID that has no viable integrity mechanism in an AI-saturated environment.