Governance News Alert: Higher Education Policy Institute (HEPI): Student Generative AI Survey 2026, and Being indispensable: Capabilities for a human-AI world, the ‘FUTURES’ framework

The third HEPI/Kortext Student Generative AI Survey report is based on responses from 1,054 full-time undergraduates in the UK.  It explores the use of AI in learning, assessment, institutional practice and personal support across higher education (HE). The second report, written by academics from Westminster University, outlines the benefits, opportunities, challenges and risks of Generative AI (GenAI) usage and argues for integration of GenAI in teaching and learning. The key points below taken from this second report are followed by “(FUTURES)”. 

The full findings of the HEPI/Kortext Student Generative AI Survey can be found at here.

The full FUTURES framework report can be found at: Being indispensable: Capabilities for a human-AI world, the ‘FUTURES’ framework - HEPI 

 At-a-glance: 

• 95 per cent of students report using AI in at least one way (compared with 66 per cent in 2024). 94 per cent say they use generative AI to help with assessed work, compared to 51 per cent the previous year (p4, p14) 
• Students increasingly utilise AI for core academic tasks, including explaining concepts, summarising academic material and structuring ideas. However, fewer students report directly generating text than in 2025 (p4)  
• Nearly two-thirds of students say assessment has changed significantly in response to AI. Some students reported that exams had got harder, as a mitigation against AI, others said AI was included in assessments (p29) 
• Some students articulate a sense of anxiety about false accusations of misconduct, although this is declining. Universities should publish clear, accessible and assessment-specific guidance on AI use (p4, p41) 
• Students from a household with the social grade of A, B or C are more likely to use some AI tools than students from a household with the social grade of C2, D or E (p10) 
• AI improves the student experience for many – but not all. Almost half of students believe AI saves time, improves understanding and provides instant support. However, a minority feel AI has worsened their experience, citing concerns about fairness, skills erosion, social isolation and future employment (p4) 
• 40 per cent of students say AI affects their sense of loneliness, while overall positive and negative impacts are reported in roughly equal measure. Around 15 per cent of students report using AI for companionship, advice or to address loneliness (p5) 
• Institutional approaches are shifting and inconsistent. Over a third of students say their institution encourages them to use AI and a similar amount say it discourages them (p5) 
• Students attending Russell Group institutions are most likely to agree that their institutions encourage the use of AI (39 per cent) – an increase from last year (26 per cent). The figure for students at pre-1992 institutions was 30 per cent (p28) 
• Two thirds of students see AI skills as essential to thrive in today’s world; but fewer than half feel their teaching staff are helping them to develop these skills for their future careers. Arts and Humanities students are particularly likely to feel under-supported (p5) 
• Universities should ensure equal access to AI, provide structured AI induction and transition support for all students, and consider revamping curricula to explicitly teach AI knowledge and skills at the general and subject-specific levels (p6) 
• Staff should have access to AI training – and be given time to invest in developing their own skills (p6) 
• Institutions should have and publicly articulate explicit institutional policies that clearly delineate acceptable and prohibited uses of GenAI in teaching, learning, assessment and professional services (FUTURES p28) 
• Institutions should establish systematic and sustained research and evaluation programmes to monitor and understand the impact of GenAI adoption on: student learning outcomes; student and staff wellbeing, workload and sense of agency; and institutional efficiency and effectiveness (FUTURES p30) 

Implications for governance: 

The HEPI survey, in partnership with Kortext, a digital HE textbook and study platform, shows that students use of GenAI, even for assessment purposes, is now almost universal. 

In tandem with this surge in usage has been a shift in how institutions view AI. The predominantly sceptical stance in the early days of the new technology is no longer evident, although not all are wholly supportive either, according to the report. Student perceptions of their institution’s attitudes paint a picture of a sector in transition, with fewer students reporting bans or restrictions on AI use. 

Governors may well be grappling with the numerous governance issues that have arisen from the advent of the new technology: from decisions about AI integration into institutions’ core activities and the resource implications of this (for instance investment in specialised software and staff training), to considerations about the student experience, teaching and learning, equity, fairness and outcomes. 

In the survey, 38 per cent of students said their institution currently provides generative AI tools for students - up from 23 per cent in 2025 and 9 per cent in 2024. By far the most common was ChatGPT, followed by Microsoft Copilot. Some respondents specified the premium version of ChatGPT. Other tools mentioned included Google Gemini and Perplexity, as well as study, writing and integrity tools such as Grammarly and Turnitin. Some respondents also listed in-house large language models developed by higher education institutions. 

While AI tools have become available at many institutions in a relatively short period, the percentage providing access is still some way from matching the half of students who want to see these tools made available. 

The report recommends structured AI induction and transition support for all undergraduates, because without it, some undergraduates risk being disadvantaged. As part of this, students should be made aware of clear expectations on the use of AI in assessment and taught a critical awareness of the limitations of AI. It is essential, says the report, that any induction focuses on “how to use AI to enhance learning, rather than outsourcing it”.   

Curricula rewrites should also be considered to explicitly teach AI knowledge and skills. At a general level, this should include core AI literacy (critical evaluation, ethical awareness and responsible use) as well as discipline-specific applications from coding and data analysis to creative practice and clinical decision making.  

Despite institutions increasingly embracing AI, the survey shows that students still feel uncertain about what is acceptable and some have concerns about being falsely accused of cheating. While all universities are likely to already have policies around expectations for assessments and academic integrity, the report suggests these need to be “clearer, accessible and assessment-specific”. From a governance point of view, issues around institutional reputation are clearly relevant. 

Students may like AI because it saves time and gives them access to round-the-clock support, but some are concerned about its impact on their skills development. The survey results demonstrate that even among students who embrace AI, individual learning experiences vary widely.  

Digital divides in attitudes and usage can also be found across subjects and groups of students. Arts and humanities students are less positive about AI tools and fear their impact on future arts careers. Better-off male students use AI more than the less affluent women. 

Striking the balance between embracing AI but maintaining students core skill and competencies, as well as fairness, is likely to be increasingly difficult as AI models becomes more sophisticated. The report suggests that running ‘AI-free’ assessments, alongside assessments that expect students to use AI, could be one answer. Institutions need to ensure they are continuing to specifically teach traditional skill sets, such as essay writing, and to emphasises to students the dangers to learning of becoming too reliant on AI. 

Governors will know that any integration of AI into curricula and/or university services is likely to generate staff training needs. According to the report, providing ongoing opportunities for teaching staff to develop discipline-specific skills in the ever-evolving world of AI is the way to “ensure that students leave their courses equipped to use AI in their future careers.” 

The FUTURES report also recommends staff training to ensure AI is integrated into disciplinary knowledge and practice and that institutions undertake a comprehensive review of existing assessment approaches to ascertain if they remain fit for purpose. 

As well as researching and monitoring AI impact on skills development, teaching and learning and student outcomes, universities also need to track the environmental impact of institutional GenAI use.  

“Monitoring and reporting mechanisms should be established to support evidence-informed decisions about when, where and how GenAI deployment delivers sufficient educational or operational benefit to justify its ultimate resource costs,” says the report.