What Higher Education Needs to Do About GenAI

Most of the public debate about generative AI in universities has focused on catching misconduct. That is necessary but not sufficient — the harder, less visible problem is that GenAI is now something most students use routinely, and institutions that treat it purely as a policing question end up with policy that lags a year behind actual practice on the ground. What the sector needs is less a bigger detection budget and more coherent decisions about curriculum, staff capability, and equitable access.

The Scale of the Gap

The scale of adoption on the student side is no longer marginal. The HEPI/Kortext Student Generative AI Survey, based on a Savanta poll of 1,041 full-time undergraduates published in February 2025, found that 92% of students now use generative AI in some form, up from 66% the year before, and that 68% believe AI skills are essential to succeed in their future careers — but only 48% feel their teaching staff are actually helping them build those skills [1]. That twenty-point gap between what students think they need and what they feel they are getting is the practical shape of the "needs" problem: it is not that GenAI is unaddressed, it is that provision has not kept pace with expectation.

A related HEPI analysis published in 2026 makes the staff-side half of that gap explicit: uneven confidence and capability among academic staff in using generative AI is producing fragile, inconsistent responses across departments — students on courses where staff feel unconfident about GenAI report more restrictive guidance, less well-designed assessment, and less clarity about what is actually permitted [2]. Institutional provision of AI tools to students has grown quickly but from a low base — up from 9% of institutions in 2024 to 23% in 2025 and 38% in 2026, according to the same survey series — which means most students are still sourcing and paying for their own tools rather than using anything the institution has vetted [1][3].

A Shared Starting Point: The Russell Group Principles

The clearest sector-wide statement of what institutions are expected to deliver came from the Russell Group in July 2023, when its 24 vice-chancellors adopted five shared principles on generative AI: supporting students and staff to become AI-literate, equipping staff to help students use GenAI tools appropriately, adapting teaching and assessment to incorporate ethical GenAI use while protecting academic rigour, ensuring equitable access to the tools themselves, and sharing practice across institutions as the technology moves [4]. Read against the adoption gap above, the principle doing the most real work is the third: teaching and assessment that assumes GenAI is available, rather than assessment designed as if it is not.

The Quality Assurance Agency's guidance points the same direction. Its Academic Integrity Charter and its "Reconsidering Assessment for the ChatGPT Era" advice both push institutions toward redesigning what is assessed rather than relying on ex-post detection of AI use [5] — a recommendation that only bites if curriculum teams have the time, training and institutional backing to actually redesign assessment at scale, which is a staff-development and workload question as much as a pedagogic one.

What Institutions Actually Need to Build

Pulling the sector guidance together, three concrete gaps recur across the QAA, Russell Group and Jisc material:

  • Staff development with real time attached. Jisc's advice to institutions explicitly warns against writing a stand-alone "AI policy" and then leaving staff to interpret it unsupported; guidance needs folding into existing assessment, data-protection and academic-integrity frameworks so it stays coherent as the tools change, and that folding-in work has to be resourced as staff time, not treated as a documentation exercise [3]. The 2026 HEPI staff-capability findings above are the direct evidence that skipping this step produces inconsistent, department-by-department outcomes for students [2].
  • Equitable, institutionally-vetted tool access. With most institutions still not providing GenAI tools directly, students who can pay for premium AI subscriptions have a working advantage over students who cannot — the opposite of what a widening-participation-conscious sector should tolerate. The Russell Group principle on equal access exists precisely because this gap was foreseeable from the outset [4].
  • Assessment redesigned for a GenAI-normal cohort, not retrofitted against it. The QAA and Jisc guidance converges on the same practical shift already visible in institutional policy responses to misconduct: staged assessment, process-visible tasks, and questions that ask for judgement rather than a free-standing artefact a model can produce whole. That redesign work is the single most resource-intensive item on this list, and the one most likely to be under-funded relative to how urgent the adoption figures above make it look [5].

Connecting to the Classroom

None of this is abstract policy work once it reaches a module leader. Our AI-Assisted Development page and Scaffolding and GenAI page work through what "assessment redesigned for a GenAI-normal cohort" actually looks like at the level of a single project brief — the same shift the sector guidance calls for, applied to specific teaching.

References

  1. Higher Education Policy Institute / Kortext, "Student Generative AI Survey 2025", HEPI Policy Note 61, February 2025. https://www.hepi.ac.uk/reports/student-generative-ai-survey-2025/
  2. Higher Education Policy Institute, "What generative AI reveals about staff capability and institutional risk in higher education", April 2026. https://www.hepi.ac.uk/2026/04/01/what-generative-ai-reveals-about-staff-capability-and-institutional-risk-in-higher-education/
  3. Jisc National Centre for AI, "Navigating the Future: Higher Education policies and guidance on generative AI", 31 July 2024. https://nationalcentreforai.jiscinvolve.org/wp/2024/07/31/navigating-the-future-higher-education-policies-and-guidance-on-generative-ai/
  4. Russell Group, "Principles on the use of generative AI tools in education", July 2023. https://www.russellgroup.ac.uk/policy/policy-briefings/principles-use-generative-ai-tools-education
  5. Quality Assurance Agency for Higher Education, "Academic Integrity Charter for UK Higher Education" and generative AI advice and resources, including "Reconsidering Assessment for the ChatGPT Era". https://www.qaa.ac.uk/sector-resources/academic-integrity