A new HEPI survey reveals that generative AI use has become 'near universal' among undergraduates within just three years, transforming from a novelty to an integral part of student life. However, students remain sharply divided on its overall impact on their learning experience. The Student Generative AI Survey 2026, conducted by Savanta with 1,054 UK undergraduates, highlights that while students are rapidly adopting AI technologies, institutional support and guidance are significantly lagging. Only 36% of students feel encouraged by their institutions to use AI, and just 38% are provided with relevant AI tools. This disparity emphasizes the critical role higher education providers must play in fostering responsible and effective AI usage, ensuring it enhances, rather than diminishes, the learning process. The report underscores the urgency for institutions to adapt their strategies to align with students' evolving AI behaviors and prepare them for a future increasingly shaped by artificial intelligence.
Key findings
The survey reveals several critical insights into student generative AI use. Firstly, AI adoption is almost universal, with 95% of students reporting use in at least one way, and 94% utilizing it for assessed work. Worryingly, the proportion of students directly incorporating AI-generated text into assessed work has increased to 12%, up from 8% in 2025 and 3% in 2024. Secondly, assessment methods have significantly changed for 65% of students in response to AI, leading to some anxiety about false accusations of misconduct. Thirdly, AI's impact on student experience is mixed: 49% believe it has improved their experience by saving time and enhancing understanding, while a minority express concerns about fairness, skill erosion, social isolation, and future employment prospects. Notably, about 15% of students are even using AI for companionship and advice to address loneliness. Finally, despite two-thirds (68%) of students deeming AI skills essential for their future, less than half (48%) feel adequately supported by their teaching staff in developing these crucial capabilities, with Arts and Humanities students feeling particularly underserved.
A polarised landscape
The 2026 survey paints a picture of a notably polarized AI landscape within higher education. This division is evident in several key areas. The proportion of students who feel encouraged by their institution to use AI (37%) is almost precisely matched by those who disagree (36%). Similarly, students are split on the social impact of AI, with 20% reporting feeling lonelier and 21% feeling less lonely. When it comes to information sourcing, there's a near-equal split between students preferring traditional sources (33%), using a balanced approach (29%), and favoring AI sources (37%). Interestingly, Russell Group institutions show a higher rate of encouraging AI use (39%), marking an increase from the previous year. Qualitative data further illustrates this dichotomy, with some students praising AI for facilitating deeper analysis and saving time on tedious tasks, while others lament, 'I’m not using my brain at all.' This highlights a critical challenge for institutions: to guide students toward leveraging AI as a tool for enhancement rather than a substitute for genuine intellectual engagement.
Key recommendations
To effectively navigate the evolving AI landscape, the report puts forth several key recommendations for higher education providers. Firstly, institutions should establish structured AI induction and transition support for all students, ensuring a foundational understanding of AI tools and responsible usage. Secondly, there is an urgent need to revamp curricula to explicitly integrate AI knowledge and skills, both generally and within specific subject disciplines, to prepare students for future careers. Thirdly, clear, accessible, and assessment-specific guidance on AI use must be published to eliminate ambiguity and promote academic integrity. Fourthly, all AI tools deemed necessary or advantageous for course content or assessment should be made universally accessible to students. Lastly, staff development is crucial, requiring access to comprehensive AI training and dedicated time for educators to develop their own AI competencies. Additionally, the Higher Education Mental Health Implementation Taskforce is urged to support research into the impact of students using AI for friendship, companionship, advice, or to combat loneliness, recognizing the emerging social dimensions of AI adoption.
