Ensuring AI literacy for all

As many as 95% of UK undergraduates now use AI in their studies, yet only 48% feel that their institutions are helping them develop the skills to use it well. This 47-point gap, noted in the Higher Education Policy Institute (HEPI) 2026 survey on generative AI[1]Student Generative Artificial Intelligence Survey, HEPI (2026)., suggests that while most institutions have an AI policy, it may not reach students where it matters most to them.

This gap prompted the Society’s Bioscience Awarding Gap (BAG) Network to facilitate the AI Literacy Equity Challenge, an interactive workshop delivered to an audience of educators and students at the Minoritised Life Scientists (MLS) Future Forum 2026 in Edinburgh. It explored whether there are inequities in how AI is used and how we can achieve greater equity in the opportunities it creates. 

The HEPI survey revealed that while AI use is almost universal among undergraduates, how it is used varies by background. Students from higher-income households are more likely than those from lower socioeconomic backgrounds to use AI for coding and data analysis, two of the most powerful applications of AI.

AI literacy is not just a question of access, but of who learns to use AI for understanding and who uses it merely to get the work done

Data from Wonkhe’s report on AI assessment, Trained to Stop Learning[2]Trained to Stop Learning report, Wonkhe, (March 2026)., reinforces this picture. It found that privately and grammar-school-educated students are nearly twice as likely as state- and further-education students to say that AI helps them learn effectively in higher education, a 32-point gap[3]‘Disadvantage is a predictor of AI use. But not in the way you might think’, Wonkhe, (March 2026).. While state school pupils were generally less anxious that not using AI would disadvantage them, privately educated students were “far more likely to have pushed the boundaries of acceptable AI use and more confident about defending what they did”. These differences show that AI literacy is not just a question of access, but of who learns to use AI for understanding and who uses it merely to get the work done.

An invisible curriculum

Many AI tools are free, but the knowledge of how to use them is typically transmitted informally through networks that not all students can access. A school student whose parent uses AI at work and discusses it over dinner arrives at university with a familiarity and confidence that no induction session can provide. This is AI literacy as cultural capital, and, like all forms of cultural capital, it is invisible until you lack it.

Meanwhile, disagreements within higher education institutions over what AI use is permitted leak directly into assessments. In the Wonkhe report, students described receiving contradictory guidance from different tutors on the same programme, or being asked to sign AI declarations they could not honestly complete. Forty-six per cent worry that not using AI puts them at a competitive disadvantage, and it is often the less experienced, less confident students who fall behind.

The Edinburgh session opened with a scenario that humanised the data. Two first-year biomedical science students face the same immunology task: a 500-word summary of the complement cascade, due the next morning. The first student, living in halls, has been shown by a cousin who recently graduated how to use the study mode of the generative AI tool to which her institution subscribes. They craft a structured prompt, and receive a clear explanation with citations and reflective questions to help them verify their output against lecture resources. The second, a working student with caring responsibilities for two children, starts after a shift at work. They are unsure whether AI is even permitted. By midnight, they have 380 uncertain words. The next day, the first student volunteers confidently during the classroom discussion. The second stays quiet, wondering whether they even belong in the classroom. 

From discussion to design

It is important to note that AI is not replacing our teaching, but it can provide access to out-of-class support that some students have always had and others have always lacked. Critically, the presence of what the Wonkhe research calls an ‘accountability moment’ – a point at which students must demonstrate understanding without AI – transforms how students use the technology, shifting it from production on autopilot to genuine learning support.

The workshop attendees went on to design a learning activity for a foundation-year biology module, considering the role AI plays in supporting (not replacing) learning, equity and access, and fair assessment. Students could use AI as a ‘study buddy’, with pre-written prompts, in-class demonstrations, peer support and an ‘exit ticket’, which serves as an accountability moment that requires an explanation of the study material without using AI. 

The AI Literacy Equity Challenge participants also drafted onboarding resources, including a toolkit of ready-made prompts at novice and intermediate levels, and an equity checklist for educators to ensure that all students can engage with AI meaningfully. 

AI is not replacing teaching, but it can provide access to out-of-class support that some students have always had and others have always lacked

The session closed with a commitment wall, on which attendees wrote one action they would take within 30 days to address AI literacy equity in their practice. The responses ranged from immediate, practical steps – “give examples of prompt structures”, “attend a study skills session on AI use” – to systemic measures: “AI policies need to be clearer within institutions, with sessions to optimise students’ ability to use AI”. Several spoke of the need to “realise that not everyone has the same AI foundation” and “build student AI confidence by facilitating guidance and support”. One delegate captured the balancing act in a single line: “AI is like medicine: take it enough, but not too much.”

The Bioscience Awarding Gap (BAG) Network is an advisory group to the Heads of University Biosciences (HUBS), an RSB Special Interest Group.

The AI Literacy Equity Challenge toolkit, including resources and worked examples from the September 2025 ‘AI for All’ event, are available here. 

This article was first published in the print issue of The Biologist, the RSB's award-winning membership magazine. Subscribe by becoming a member today.

References