Rigour and warmth are not a trade-off: academic conduct needs less policing and more care

Join the Policy Institute at King’s College London and HEPI on Tuesday 15 September, from 6.30pm to 7.45pm, at Bush House, London, for a free event exploring the future of higher education in England. Drawing on Professor Sir Chris Husbands’ report, New choices: Revisiting futures for higher education in England, the discussion will consider how financial pressures, generative AI, changing student expectations and declining public trust are reshaping the sector – and the choices universities and government need to make now to secure a sustainable future. Register here.

This blog was kindly authored by Dr Chelle Oldham, The Open University & University of Glasgow, University Academic Integrity Co-Lead.

Generative AI has done something to academic conduct that no policy review ever managed: it has made the caseload impossible to ignore – referrals are up across the sector and detection tools are flagging work at volume. The institutional reflex under that pressure, understandable but wrong, is to police harder. Tighten the rules, catch more, move faster. I want to suggest the opposite. The conduct conversation is one of the few moments where a university decides, in practice, what kind of institution it is, and the evidence of my own caseload is that we get fairer, more defensible and more honest outcomes when we lead with compassion rather than suspicion.

Compassion here is not the soft option, and it is emphatically not about lowering standards. It means holding students to high standards while assuming good faith and giving them the support to meet them. Rigour without warmth produces a process that is correct and unjust at the same time; warmth without rigour helps no one. The skill (and it is a skill, not a personality trait) is holding both. In practice, that starts with curiosity rather than accusation: when something looks wrong, the most useful opening is a question, not a verdict. Most of what surfaces turns out to be a gap in learning, not dishonesty: often a student leaning on assistive tools they were entitled to use, who never knew where the line sat.

This matters more sharply in the AI era because of who gets flagged. Detection tools do not catch dishonesty; they catch difference. The students most likely to be marked out by an unusual style or a probability score are disproportionately neurodivergent, disabled, or writing in English as an additional language. In 2026, we are investigating students for using text-to-speech software, spelling tools and AI-assisted writing aids that were, in many cases, recommended to them through their Disabled Students’ Allowance. We are, in effect, policing the very accommodations we prescribed. A growing body of research confirms far higher false-positive rates for exactly these students, penalising those who are not cheating on suspicion alone. A flag is a reason to ask a question. It is never, on its own, a conclusion. Treating it as proof does not just risk an individual injustice; it builds a process that systematically disadvantages the very students our equality duties exist to protect. As Sarah Eaton puts it, ‘there can be no integrity without equity’.

Conduct work can also land at one of the hardest moments in a student’s life, when the allegation is the least of what someone is carrying. The line I hold onto is that a conduct officer is not a clinician and does not need to be. The role is to run a fair process with care and to connect the student to people who can help, not to diagnose or counsel. That clarity is liberating. It means signposting wellbeing support early, making the process itself accessible, and being willing to pause when a student is too unwell to engage fairly. You can take an alleged breach seriously and treat the person with care at the same time; these are not in tension.

The sector now has a hard reason to take this seriously as well as a moral one. The Abrahart litigation has made plain that a university’s duty of care and its anticipatory equality duties are not aspirational language in a strategy document. They are enforceable, and they apply to the processes we run as much as to the teaching we deliver. A conduct process designed only for the confident, well and articulate student is not a defensible process. It is a liability that happens to look like rigour.

The constructive alternative is not more surveillance but better education, delivered earlier. The Educate–Enable–Expect model I have argued for elsewhere starts from the premise that most students want to do the right thing and that our job is to make sure they know how: educate them about what integrity means in practice, including how their assistive tools interact with assessment; enable them to practise in formative spaces before the high-stakes moment; and only then expect them to take ownership. For disabled students navigating already-complex accommodation landscapes, that sequence is not just pedagogically sound: it is the fair one.

None of this requires new policy. It requires us to stop treating compassion and rigour as a trade-off and to build the alternative into how we train and support conduct officers from the start. Difference is not dishonesty. A flag is not a finding. And a fair process, run with care, is the most rigorous one we have.