Over November and December 2025, Jisc convened three roundtables: two face-to-face, in Manchester and London and one online, aimed at staff from Scottish institutions. Bringing together 71 colleagues from across 59 tertiary institutions to explore what skills, knowledge and capabilities learners and students will need in an AI-enabled world. Each session was grounded in the Policy Connect briefing ‘Skills in the Age of AI’, using its themes to examine how institutions are responding to rapid technological change. The roundtables set out to surface regional insight, identify shared challenges, and build an evidence base to inform a national position statement on skills in an AI age.
Although each region had its own character, the conversations revealed shared pressures, common concerns and a clear picture of what the sector needs next.
The UK faces a foundational digital and AI literacy gap
In every region, participants noted that their institutions consistently overestimate students’ baseline digital literacy. Many learners and students struggle with basic tasks such as formatting documents, evaluating online information or navigating digital tools. Staff confidence varies widely too, leaving learners and students without consistent guidance.
This gap makes AI literacy much harder to build, including areas such as:
- understanding what AI is, what it isn’t, and how it works
- distinguishing when AI should and should not be used
- strong information literacy to navigate bias, misinformation and deepfakes
- ethical and environmental understanding
- awareness of privacy, risk and terms of use
- the confidence to challenge AI outputs rather than accept them
The myth of the ‘digital native’ was identified as a persistent barrier. Young people are fluent in social media yet often lack the digital grounding needed for academic or workplace AI.
Equity concerns surfaced repeatedly. Learners and students who can afford premium tools gain an advantage, while others rely on outdated or limited resources. International and neurodiverse learners and students are sometimes wrongly accused of AI misuse because their writing patterns differ from expected norms. Learners and students regularly upload sensitive information to unlicensed tools without understanding the risks.
Regional insights
Attendees at the Manchester roundtable reflected the most acute digital readiness challenges, especially within FE, where staff routinely need to teach foundational digital skills before addressing AI. Manchester attendees also noted that learners’ lack of confidence often drives unskilful AI use or overreliance as a shortcut.
London attendees highlighted sharp differences in starting points tied to socioeconomic background, schooling and access to devices, and raised concerns about ‘hidden AI’ in everyday tools that learners may not recognise.
Attendees at the Scottish roundtable stressed the importance of informed choice, including the right to opt out of AI use on ethical or personal grounds. Scotland’s attendees also described a three-layer view of AI literacy, spanning context, learning and discipline-specific workplace use.
Across the UK, colleagues agreed that AI literacy must be lifelong, life-wide and embedded, not treated as an optional add-on.
Human skills are now the primary differentiator
Despite regional differences, the UK-wide consensus was striking: human skills matter more, not less, in an AI world.
Participants across the three roundtables highlighted similar essential capabilities to the Policy Connect Skills in the Age of AI briefing: curiosity, adaptability, resilience, collaboration, creativity, critical thinking, reflection, empathy and judgement.
These human attributes allow learners and students to navigate uncertainty, challenge AI outputs and make informed decisions about when to rely on technology. They also protect against deskilling, overreliance and the loss of intellectual confidence.
Regional insights
Attendees at the Scotland roundtable foregrounded informed choice, rights-based digital practice and a continued need for human alternatives in areas such as education, welfare and healthcare.
In Manchester, attendees emphasised rebuilding confidence and curiosity after years of learners and students being over-directed, and highlighted the importance of resilience and acceptance of failure as part of learning.
In London the social justice dimension was highlighted, arguing that human capability development must sit alongside technical skills to avoid widening inequality, and stressed the growing importance of criticality in understanding the agendas behind AI systems.
The message was clear: human and AI capabilities must evolve together.
Current assessment models are no longer fit for purpose
Assessment emerged as the most urgent area requiring reform. Across all regions, participants agreed that essays and knowledge-recall tasks reward polished written output rather than genuine understanding. These formats also encourage learners and students to use AI in hidden or unskilled ways.
Key concerns included:
- widespread confusion about acceptable AI use
- learners and students are afraid to declare AI use, even when it is responsible
- inconsistent rules within and across institutions
- awarding bodies restricting innovation, particularly in FE
- assessments that feel disconnected from employability
- multilingual and neurodiverse learners and students being penalised unfairly
- increasing student use of AI for emotional support without adequate guidance
Staff described significant anxiety about distinguishing responsible use from misconduct. Traffic light rules fail to reflect real human–AI interaction, and staff often lack time to explore alternatives.
Participants saw assessment reform as essential and suggested:
- programme-level and holistic approaches
- capstone tasks
- multimodal and authentic assessments
- selective supervised checkpoints where needed
- clear two-lane models that separate assessment for learning from assessment of learning (University of Sydney)
- explicit assessment of process, reflection and critique of AI outputs
Positive examples were shared, such as learners and students working with tutor-designed chatbots to support website projects, generating strong engagement and skill development.
Regional insights
Manchester’s attendees voiced the sharpest frustration with awarding bodies and legacy assessment models, particularly in FE where reform is tightly constrained, and noted that learners rushing to ‘catch up’ often turn to AI in risky ways.
In London there was a focus on trust, transparency and complexity, including how to distinguish legitimate assistance from misconduct, and raised the emerging issue of learners being accused unfairly due to linguistic or cultural differences.
In Scotland, attendees stressed fairness, transparency and choice, including for learners who opt out of AI use, and emphasised the need for shared vocabulary and national coherence.
Despite pockets of innovation, system conditions continue to block meaningful reform. Assessment reform cannot be delivered institution by institution. Awarding bodies must be part of the solution.
Lack of time is the overriding barrier to progress
The strongest and most consistent message from all three roundtables was that staff do not have the time required to respond to AI meaningfully. Staff in FE, HE and adult learning are already stretched, with no capacity to explore AI tools, redesign assessment or support learners and students in using AI safely.
Participants described:
- staff working without protected time for development
- staff under intense qualification-driven pressure
- staff juggling heavy teaching loads with REF expectations
- tutors needing to teach basic digital skills before tackling AI
- inconsistent or unclear policies adding to workload
- no institutional space for reflection, testing or collaboration
The discussion in Manchester and Scotland also highlighted that staff feel caught between conflicting pressures: to innovate, to stay compliant, to maintain student trust and to meet performance expectations.
The message was unequivocal: the sector is willing but not resourced. AI capability cannot be built on top of an overloaded system. Staff development was seen as the single most important national need.
Learner and student trust, wellbeing and ethical concerns are rising
A growing group of learners and students are choosing not to use AI due to ethical objections, environmental concerns, fear of dependency, mental health worries or mistrust of data handling. Others rely on AI in ways that bypass learning or expose personal information.
Regional insights
In London we heard the strongest safeguarding concerns, including learners using AI for emotional support, risks linked to deepfakes, voice clones and synthetic identities, and poorly governed edtech partnerships that may undermine safety or integrity.
Manchester attendees highlighted that learners’ fear of getting things wrong is pushing some toward over-reliance on AI, particularly among adult learners trying to keep pace.
In Scotland, the discussion emphasised rights-based questions about where AI should not be used, highlighting informed choice, the right to decline AI, and the need to ensure human alternatives remain available in education and support services. Participants emphasised the importance of transparency in how AI operates, the use of data, and its limitations. This aligns with Scotland’s wider rights-based digital policy approach.
Participants agreed that AI literacy must include personal risk awareness, not just academic or employability considerations. The sector needs shared frameworks for AI safety, trust and wellbeing.
What the UK tertiary sector needs next
Participants across all three regions argued for coordinated national action to shape AI adoption in education. Key priorities include:
- a national baseline for digital and AI literacy, so learners and students arrive with consistent foundations rather than wide disparities
- clear definitions of responsible and acceptable AI use in assessment, removing ambiguity for staff and learners/students and reducing misconduct disputes
- modernised assessment structures, for example, in HE, enabling programme-level, and in FE, support from awarding organisations in implementing authentic and process-focused assessment that AI cannot simply replace.
- develop tailored guidance for FE and HE, building on DfE’s 2025 guidance[1], adapting safeguarding and cybersecurity for adult learners, higher education context and workplace-readiness rather than school-level safeguarding.
- extend DfE frameworks to cover not only under-18 learners, but also adult learners, international and neurodiverse students, ensuring policies cover data, bias, misinformation, identity misuse, consent, and deepfake risk across the entire tertiary sector.
- sustained, high-quality staff development, with protected time and shared resources so staff can engage confidently with AI and redesign their practice
- equity-focused investment, ensuring learners and learners and students are not disadvantaged by their ability to pay for premium tools or by uneven institutional provision
- consistent institutional approach to policy and guidance, co-developed by academic and professional services so staff and learners/students receive clear and unified guidance
- greater alignment across schools, FE, HE and employers, reducing the gaps between sectors and helping learners and students transition smoothly
- a consistent, sector-wide AI procurement and due-diligence framework that builds on the DfE’s Product Safety Expectations[2], ensuring these protections apply across FE, HE and adult learning. Vendors should be required to evidence transparency, bias testing, data use and system limitations before their tools are deployed in tertiary education
- coherent approach to edtech partnerships, ensuring commercial AI products align with educational values, protect learner data and avoid dependency or hidden risks
Participants were clear that none of this can be achieved by institutions working alone.
National coordination, shared standards and sector-wide leadership are essential for safe, equitable and future-ready adoption of AI.
Conclusion
The UK tertiary education sector is facing a pivotal moment. Preparing learners and students for an AI-enabled world requires far more than teaching tool proficiency. It demands addressing foundational digital literacy, rethinking assessment, investing in staff capability, strengthening safeguarding and wellbeing, and aligning national policy with the realities of practice.
Across our Scotland, Manchester and London workshops, the message was united: AI must be understood, taught and governed through a human-centred, ethical and future-focused lens.
Insights from these roundtables will directly inform Jisc’s Artificial Intelligence future activities.
Thanks to Participating Institutions
Abertay University
Barnsley College
Basingstoke College of Technology
Bath Spa University
City College Norwich
City of Glasgow College
Colchester Institute
Coleg Cambria
Coleg Sir Gar
Coventry College
Edinburgh College
Edinburgh Napier University
Forth Valley College
Glasgow Caledonian University
Harper Adams University
Heriot Watt University
Imperial College London
Keele University
King’s College London
Liverpool School of Tropical Medicine
London College of Communication
Milton Keynes College
North East Surrey College of Technology
Queen Margaret University
Queen Mary University London
Queen’s University Belfast
Regent’s University London
Robert Gordon University
Scotland’s Rural College
The London School of Theology
The Open University
Ulster University
University of Aberdeen
University of Bath
University of Birmingham
University of Bristol
University of Central Lancashire
University of Derby
University of Dundee
University of Exeter
University of Glasgow
University of Hertfordshire
University of Leeds
University of Liverpool
University of London
University of Manchester
University of Portsmouth
University of Roehampton
University of Southampton
University of St Andrews
University of Strathclyde
University of Sunderland in London
University of the Arts London
University of the Highlands and Islands
University of Wales Trinity Saint David
University of Winchester
University of York
Westminster Adult Education Service
Woking College
World Skills UK
[1] Generative artificial intelligence (AI) in education, Updated 12 August 2025 and Meeting digital and technology standards in schools and colleges Updated 17 November 2025
[2] Generative AI: product safety expectations Published 22 January 2025
Find out more by visiting our Artificial Intelligence page to explore publications and resources, learn more about our communities and sign up for our AI Literacy training.
For regular updates from the team sign up to our mailing list.
Get in touch with the team directly at AI@jisc.ac.uk