Advance HE's Assessment Forum says AI-era assessment needs a clearer student voice evidence trail
Updated Jul 29, 2026
Advance HE's latest assessment forum matters because it links AI policy, assessment design, and student trust in a way universities can act on now. On 23 July 2026, Advance HE published Choices, not policing: reshaping assessment for the AI era, summarising a Manchester forum of more than 50 senior academic leaders who discussed trust, the 2028 regulatory horizon, and the future of teaching and learning. For teams responsible for student voice, the practical takeaway is that AI-era assessment change now needs a clearer evidence trail from students on whether briefs, permissions, and feedback still make sense in practice.
What has changed in AI-era assessment practice
The immediate development is not a new OfS rule. It is a stronger sector signal that assessment redesign now needs a more explicit, reviewable operating model. Advance HE says the forum was built around three provocations: the cost of trust, the 2028 regulatory horizon, and the digital education horizon. Delegates captured their own observations during the day, and Advance HE says a short paper drawing on those discussions will follow. That makes this a practice and governance development rather than a formal policy announcement, but it is still a clear sign of where the sector conversation has moved.
Advance HE reports that the first panel rejected a simple return to closed-book exams. It highlights Professor Judy Williams describing a staged model at Queen's University Belfast, with clear points where AI use is encouraged, controlled, or restricted. It also points to Bristol's move towards programme-level assessment to reduce overall burden and rebuild more human-centred teaching. The common thread is that assessment security should be proportionate to what is being judged, and educational design should do more of the work.
"an 1826 answer to a 2026 problem"
That line came in the panel on the 2028 regulatory horizon, where Jo Coward warned against a reflex return to invigilated, on-premise assessment. Advance HE says panellists argued that if the sector does not respond coherently, the OfS could be pushed towards new conditions requiring auditable AI assessment policies. In the final panel, contributors argued for discipline-specific AI literacy, more equitable access to tools, interactive rubrics, and stronger student trust in institutional intentions around permitted use. This is a UK-wide enhancement signal rather than a new regulation, but it is directly relevant to how universities explain assessment change and collect evidence on whether students understand it.
What this means for institutions
First, universities should now treat AI-era assessment as an assurance problem as well as a design problem. The forum discussion suggests that broad institutional AI principles are no longer enough if modules are still varying in what they allow, how they phrase it, and how feedback is meant to help students adapt. Institutions should therefore review assessment briefs, acknowledgement expectations, rubric design, and follow-up questions together rather than in separate workstreams. That sits closely beside the recent QAA warning about inconsistent AI practice becoming a student experience risk.
Second, the practical evidence burden is rising even before any new regulation appears. If senior teams want to show that assessment redesign is fair, proportionate, and understood by students, they need feedback routes that test clarity, consistency, workload, and trust at programme level. A student comment analysis governance checklist is useful here because it forces teams to define which routes collect that evidence, who reviews it, and how mismatches between modules are escalated. The benefit is not more data collection for its own sake. It is a clearer line from local student experience to a decision senior leaders can defend.
Third, programme-level change matters more than isolated pilots. Advance HE's examples point to staged models, burden reduction, and more human-centred teaching rather than one-off AI bans. For Student Experience teams, PVCs, and quality professionals, the operational takeaway is straightforward: compare student responses across modules that sit within the same programme, and look for drift in permissions, rationale, and feedback usefulness before the next assessment cycle starts. That is often where student confusion appears first.
How student feedback analysis connects
This is where open-text feedback becomes more useful than headline sentiment alone. Students will usually tell you exactly where the new model broke down: one module allowed brainstorming but another stayed silent; a rubric mentioned process but feedback still rewarded polished output; a viva felt clearer than an essay brief; or AI acknowledgement rules arrived too late to shape behaviour. Those are the details institutions need if AI-era assessment is going to be reviewable rather than rhetorical.
A structured NSS open-text analysis methodology helps teams compare those comments across module evaluations, in-term pulse work, NSS, and representative channels without treating every AI-related issue as the same problem. Student Voice Analytics is one restrained way to support that kind of cross-survey review. The broader point is methodological: if universities may soon be expected to evidence how AI-era assessment is working, the comment-analysis process also needs to be stable enough to defend.
FAQ
Q: What should institutions do now if they are redesigning assessment for AI use?
A: Start by mapping where AI permissions, restrictions, and feedback expectations currently differ across a programme. Then add a small number of targeted prompts to existing module evaluation, rep, or pulse routes so you can test whether students understood the rationale, the rules, and the intended use of feedback. The goal is to surface drift early enough to correct it.
Q: Does Advance HE's forum announce a 2028 rule change?
A: No. Advance HE published the forum summary on 23 July 2026, and the "2028 regulatory horizon" appears as a panel discussion about where pressure could move if the sector does not respond coherently. Advance HE says a short paper from the event will follow. The scope is UK higher education practice and assurance, not a formal regulatory decision.
Q: What is the broader implication for student voice?
A: Student voice is moving closer to assessment assurance rather than sitting only in post-hoc enhancement work. Institutions increasingly need evidence not just that students had an opinion, but where rules, burden, and feedback quality changed across a programme and whether those changes were actually understood.
References
[Advance HE]: "Choices, not policing: reshaping assessment for the AI era" Published: 2026-07-23
Request a walkthrough
Book a free Student Voice Analytics demo
See all-comment coverage, sector benchmarks, and reporting designed for OfS quality and NSS requirements.
- All-comment coverage with HE-tuned taxonomy and sentiment.
- Versioned outputs with TEF-ready reporting.
- Benchmarks and BI-ready exports for boards and Senate.
UK-hosted · No public LLM APIs · Same-day turnaround