By Dr. Elizabeth Pinneh,
Picture a student on their first day at university. They have cleared every hurdle the admissions process placed before them: personal statement, finance application, and induction week. What nobody told them is that, quietly, a different race has already begun. Around them, peers from more privileged backgrounds have arrived fluent in AI tools: synthesising readings, refining essays, preparing for interviews. They simply arrived knowing, because knowing, in this landscape, is something that money, networks and prior exposure affords.
This is the new face of educational inequality. It does not announce itself through a rejected application or an empty seat in a lecture hall. It arrives through the digital back door. If the widening participation mission means anything, it must now grapple with a truth that is both urgent and underacknowledged: access to university is no longer enough if students do not have equitable access to the digital knowledge, tools and confidence required to succeed. The divide is already here, and it is widening. The evidence makes for uncomfortable reading. The Higher Education Policy Institute’s Student Generative AI Survey 2026 found that AI use among undergraduates is now near-universal, with 95% of students reporting using AI in some form and 94% using generative AI for assessed work (HEPI, 2026). Despite this saturation of use, just 48% of students felt that their teaching staff were actively helping them develop AI skills for their future careers (HEPI, 2026).
The gap between adoption and guided, equitable development is not closing; it is growing. Digital Experience Insights Survey 2024/25 makes the structural dimension of this inequality visible. Black/Black British and Asian/Asian British students were disproportionately likely to lack suitable devices, with 44% and 47% respectively reporting problems from inadequate equipment (Jisc, 2025). Connectivity issues, the cost of premium AI tools and insufficient institutional guidance compound one another into a layered disadvantage that is simultaneously material and cultural. The annual cost of a full suite of premium AI tools and educational plug-ins can reach approximately £1,000 per year (Jisc, 2025). A figure that renders meaningful AI engagement an aspiration rather than a baseline for those on the tightest budgets.
This inequality does not emerge in a vacuum. It maps onto, and reproduces, long-established patterns of structural disadvantage. Students from lower socioeconomic backgrounds, minority ethnic groups, mature learners and disabled students face barriers that are tangible such as inadequate devices, unaffordable subscriptions. Intangible barriers are a deficit of confidence, the absence of peer networks through which digital literacy is informally transmitted, and a diminished sense that advanced technological tools are for people like them. The instinctive institutional response to a skills gap is to create a workshop. This misdiagnoses the problem. Widening participation research is consistent: optional enrichment provision is disproportionately accessed by students who are already advantaged. The students who most need structured AI literacy support are the least likely to self-select into drop-in sessions, particularly when they are also managing part-time employment, caring responsibilities and the cognitive burden of being the first in their family to navigate higher education.
The rapid adoption of AI across academic and professional life is creating a new layer of exclusion. Those who understand how to use AI productively, critically and ethically will gain measurable academic and professional advantage. Those who do not, will be doubly disadvantaged. First in their studies, then in a labour market that increasingly assumes AI fluency. Digital equity is therefore not a satellite concern. It is central to student success, belonging, retention and graduate outcomes, and it demands the same policy coherence that institutions bring to attainment gaps and fair access.
From AI Framework to Action: Literacy, Equity, and What Universities Must Do
UNESCO’s AI Competency Framework for Teachers (UNESCO, 2024a) provides a rigorous model for institutional action. It identifies the competencies educators require to use AI responsibly, encompassing curricular integration, learning objective design and an explicit grounding in ethics and human rights. Adapted for the UK higher education context and accompanied by funded staff development, it provides a principled basis for ensuring inclusive AI pedagogy becomes a consistent expectation rather than an individual enthusiasm. The complementary AI Competency Framework for Students (UNESCO, 2024b), organised across four dimensions: human-centred mindset, ethics of AI, AI techniques and applications, and AI system design, positions AI literacy as a matter of critical agency and ethical reasoning, not mere tool proficiency. Students equipped with these competencies can interrogate the bias of AI outputs, make informed judgements about appropriate use and engage with AI as reflective practitioners. This is precisely the literacy that widening participation students are least likely to receive systematically and most likely to need.
The OECD’s AI Exposure Measure (OECD, 2026) maps AI capability indicators to occupational requirements across nine domains including reasoning, social interaction and metacognitive capability. Its key finding is instructive: while routine information-processing roles face the highest AI substitution risk, occupations demanding complex judgement, contextual reasoning and ethical decision-making exhibit the largest gaps between current AI capabilities and human workers. The jobs least susceptible to displacement are precisely those requiring the distinctly human capacities: critical analysis, relational intelligence and ethical awareness. The goal, therefore, is not to train students to use AI tools. It is to develop the higher-order capabilities that make graduates indispensable alongside AI. As employers increasingly structure roles, around data literacy and AI-enabled workflows, universities have a direct responsibility to ensure that students from underrepresented backgrounds are not excluded from the future labour market by a skills gap that institutional design could prevent (OECD, 2026).
The response this analysis demands is structural, not supplementary. This begins at induction. A short diagnostic audit, embedded in registration, should cover device access, home connectivity, prior AI experience and confidence with digital learning environments, with responses disaggregated by widening participation characteristics such as first-generation status, disability and care-experienced background. That data must trigger proportionate action: device lending schemes, institution-wide AI platform access, and a scaffolded AI literacy module in induction week, framed as standard academic preparation rather than remediation. Assessment design is where structural commitment becomes measurable: process-focused, authentic tasks, portfolios, reflective commentary and staged drafting with documented reasoning, reduce the advantage that premium AI access confers and reward the critical, contextualised capabilities the labour market most values (OECD, 2026).
The student who arrives at university without AI literacy is not simply behind on a technical skill. They are entering an institution whose informal curriculum is increasingly mediated by tools they have never been taught to use, question or interrogate. To leave that gap unaddressed is not a neutral choice. It is a decision, however unintentional, to reproduce inequality by other means. Closing it is not a concession to technological change. It is the widening participation mission, restated for the age of AI.
References
HEPI (2026) Student Generative AI Survey 2026, Report 199. London: Higher Education Policy Institute. Available at: https://www.hepi.ac.uk/wp-content/uploads/2026/03/HEPI-Report-199-Gen-AI-Survey-2026.pdf
Jisc (2025) Student digital experience insights survey 2024/25: UK higher education (HE) survey findings. Bristol: Jisc. Available at: https://repository.jisc.ac.uk/10242/1/DEI-2025-student-he-report.pdf
OECD 2026,The OECD AI exposure measure: Mapping the OECD AI capability Indicators to occupations, OECD Artificial Intelligence Papers, no. 59, OECD Publishing, Paris. Available at: https://www.oecd.org/content/dam/oecd/en/publications/reports/2026/05/the-oecd-ai-exposure-measure_489cfd42/f3da0f0a-en.pdf
UNESCO (2024a) AI Competency Framework for Teachers. Paris: UNESCO. Available at: https://unesdoc.unesco.org/ark:/48223/pf0000391104/PDF/391104eng.pdf.multiUNESCO (2024b) AI Competency Framework for Students. Paris:
UNESCO (2024b) AI Competency Framework for Students. Paris: UNESCO. Available at: https://unesdoc.unesco.org/ark:/48223/pf0000391105/PDF/391105eng.pdf.multi