Written evidence by Professor Gunter Saunders and Dr Doug Specht (AIE0065)
Education Committee
The use of Artificial Intelligence and EdTech in Education inquiry
This submission explains that generative artificial intelligence (GenAI) could transform higher education, but these benefits will only be realised if institutions act quickly rather than follow the usual slow pace of technology adoption in UK universities.
We highlight two main messages in the evidence below:
The authors recently published a report with the Higher Education Policy Institute (HEPI) called Being Indispensable: Capabilities for Human-AI World, the Future's Framework (HEPI Report 198, March 2026). This report offers a practical framework for achieving the two goals above. This submission uses evidence and recommendations from that work to address the committee's questions.
About the authors.
Dr Doug Specht leads the Westminster School of Media and Communication at the University of Westminster. Professor Gunter Saunders is Director of Digital Capability Development and AI Leadership at the same university. Together, we have extensive experience researching and guiding how universities adopt generative AI in teaching and learning. Our recent HEPI report (Report 198, 2026) forms the evidence base for this submission and introduces the Futures Framework, a practical model for developing the seven key human capabilities graduates need in an AI-enabled world.
We are submitting this evidence because we believe the Committee's inquiry is addressing one of the most urgent and life-changing policy questions facing education today. Our recent research and institutional experience have given us direct insight into both the opportunities and risks of GenAI in higher education and into the gap between where the technology is and where institutional policy and practice currently sits. This reflects a growing structural misalignment across the sector: while institutional policy development proceeds through formal governance processes, staff and students have already established their own day-to-day practises with Gen-AI tools. The result is not simply a lag in policy but a divergence between formal regulation and lived educational practise that must be addressed directly.
We are concerned that without clear direction and mandates from government, sector bodies, and the Committee, institutions will not change quickly enough to meet the challenge. Slow responses are already affecting students the most.
1. The current situation is urgent. Higher education cannot afford to move slowly.
Higher education institutions are used to taking time to make decisions, which reflects the proud traditions of the sector. However, changes to curriculum, policy, and governance often take years instead of months. While this careful approach can be a strength, it is now a serious drawback. GenAI tools are advancing quickly, and universities are struggling to keep up. Evidence from schools, colleges, and universities shows that students and staff are already using these tools widely, often before any official guidance or policy is in place.
Research shows that about three-quarters of young people aged 13 to 18 have already used generative AI, and similar trends are seen in universities. Staff and students are quickly adapting to these new tools, often without clear institutional guidance. The debate must now focus on how to integrate GenAI responsibly and safely, not whether it should be used. Institutions that do not provide timely and practical answers risk falling behind and leaving their students behind as well.
The committee should consider ways for the government to encourage, and if needed require, higher education institutions to create and publish clear GenAI strategies with set timelines. Voluntary action will not be enough. This challenge needs more than just regulation; the main goal is to move from simply regulating AI to meaningfully integrating it into teaching, learning, and assessment.
1.1 The opportunities:
When used carefully, GenAI can greatly improve higher education. These tools offer real and important opportunities.
These opportunities are real and already happening in universities that have acted quickly and with purpose. The key question for the committee is how to make sure the benefits of GenAI are shared fairly across all institutions, not just those with more resources or authority to act first.
1.2 The risks:
Without proper oversight, GenAI can increase inequality and harm learning. Poorly managed GenAI adoption carries serious risks. Our research and other evidence highlight several related concerns that need urgent attention:
Deepening inequality
Students who can pay for premium GenAI tools will have access to much better systems than those using free versions. Without action, GenAI could make existing divides in higher education worse. Government and sector bodies should ensure fair access to high-quality tools, possibly through collective licensing. Institutions should also monitor and report on access equity as a standard practice.
Embedded bias
The large language models behind GenAI tools can show biases from their training data. Students and staff need to learn how to spot and question these biases. This is not just a technical issue; it is also an educational and ethical one, so it should be built into the curriculum rather than added as extra training.
De-skilling and over-reliance
If GenAI is used without critical thinking, it can weaken deep learning, independent thought, and real academic skills. The committee is right to worry about students losing important abilities. Banning GenAI is not the answer, as it is both unrealistic and unhelpful. Instead, GenAI should be integrated in ways that require students to think critically, assess its outputs, and build their own unique human skills. These inequalities are not just about money, but also about differences between institutions and countries. Differences in infrastructure, investment, and access to premium AI tools could create big gaps in skills within the UK and worldwide, making educational inequalities worse.
Academic integrity
The wide use of GenAI has changed how assessments work. Traditional coursework and unsupervised assessments are no longer suitable as they are. Assessments need to focus on skills that require human judgement, such as interpretation, application, ethical reasoning, and creativity. Institutions should review their assessment practices thoroughly. They should also reconsider what academic integrity means today and whether current definitions match how people work in the 21st century.
Environmental Impact
GenAI infrastructure uses a lot of energy and water, and higher education does not fully understand these impacts. Institutions should include environmental factors in their AI purchasing decisions. The government should help sector bodies create shared guidance and collective strategies for sustainable GenAI use.
Fragmented institutional governance
Across institutions, GenAI is being adopted in a scattered and uncoordinated way, often led by local efforts instead of a clear strategy. Many staff are likely using public GenAI tools for their work instead of secure, institution-approved systems. This creates risks such as inconsistent data management, limited checks for bias, and unclear responsibility for AI-assisted work. Without strong oversight, these risks will likely grow, especially since universities handle sensitive student data.
2. Developing human capabilities: The Futures Framework
Our main argument, explained in detail in the HEPI report, is that higher education needs two things: responsible integration of GenAI and a systematic focus on developing the human skills that will still matter in an AI-rich world. Current competency frameworks are often either too broad or too technical, making them hard to use in academic programs.
National and international frameworks have made valuable contributions, but they are often either too strategic to apply in daily teaching or too focused on technical skills, missing the full range of abilities needed across different subjects.
The Futures Framework we have developed addresses this gap. It proposes seven interconnected domains of indispensable human capability:
Each domain is set up as a progression from basic to advanced skills, with clear examples that students and institutions can use to track growth. The framework should be built into courses and assessments, not treated as a separate module, and it works for all subjects, including the humanities, where GenAI’s impact is often overlooked. We suggest using the Futures Framework together with the JISC AI Maturity Toolkit, which focuses on infrastructure and governance. These two frameworks together cover both safe AI deployment and the development of human skills needed to use AI well.
3. Implications for teaching practice, curriculum and assessment
GenAI will have a major impact on teaching, and this is not yet fully recognised. Integrating GenAI effectively requires not just technical changes, but also important shifts in professional habits and culture.
Staff need confidence and practical support to experiment with AI within their own disciplinary context, to articulate clear expectations to students, and to redesign assessments that remain meaningful.
Professional development for academic and support staff should be a real institutional priority, not just optional training. It should focus on three main areas: using GenAI tools critically and creatively; ethical practice and compliance; and working across disciplines, with extra support for the humanities where issues of authenticity and interpretation are especially important.
GenAI literacy should not be taught separately from subject knowledge. Students want clear guidance on expectations, fairness, and how to build confidence in their own abilities when working with AI. GenAI literacy needs to be built into all programs, so students learn to use GenAI critically in their field, judge when AI-generated content is useful, and develop digital skills specific to their discipline. The national curriculum review should address this integration directly.
When it comes to assessment, banning GenAI is not effective and should not be the default. Instead, assessments should be redesigned so students show they can evaluate, critique, and responsibly use GenAI in their own work. The main question is not whether students can work without AI, but whether they can think critically about AI outputs and use them responsibly. This is a true and challenging test of graduate skills.
4. Government's role: Framework, Regulation and Investment.
The committee asks if the government has a clear enough framework for guiding and regulating GenAI in education. We believe the current framework is not enough for the speed of change. We offer these specific recommendations for government action:
Institutional Policies
Government should require higher education institutions to publish clear and detailed GenAI policies, covering acceptable use, assessment, academic integrity, and fair access, within a set time frame. Generic statements are not enough; each institution needs clear policies suited to their subjects. Because universities handle large amounts of data and make important decisions, they must be supported to act responsibly and ethically in their use of AI. Governance should balance safety and responsibility with the need to make the most of GenAI.
Equitable Access
Sector bodies and government need to set up ways to ensure all students, especially those from disadvantaged backgrounds, have fair access to high-quality GenAI tools and infrastructure. Government investment in digital infrastructure should include GenAI access as part of digital equity.
Research and Evaluation
Government should fund ongoing research and evaluation to track how GenAI affects student learning, staff and student well-being, and how well institutions work. GenAI is being adopted faster than evidence is being collected, so a coordinated, sector-wide evaluation is needed.
Sector-wide and international collaboration
The challenges of GenAI are not limited to single institutions or countries. Government should support ways for the sector to coordinate and share knowledge internationally, so institutions can learn from each other and avoid fragmented or duplicated efforts.
The Environment
Any national framework for AI in education should include environmental sustainability. The energy and resource use of widespread GenAI adoption needs clear oversight.
National Curriculum and Assessment Review
The national curriculum and assessment review should clearly address how students can develop the skills needed for an AI-enabled workplace across all subjects and stages. The Futures Framework is one practical way to approach this development.
5. Conclusion
Generative AI is already changing higher education. It can bring many benefits, such as personalised learning, wider access, less administrative work, and better assessments—but only if institutions act quickly and with clear purpose. If action is too slow or cautious, the risks are real: greater inequality, weaker academic integrity, and students who rely on AI without building the essential human skills they need for work and civic life.
Higher education is used to moving at a careful, slow pace, but this approach does not fit the current situation. The committee now has a key chance to show government, institutions, and the sector that change must happen faster, and to support the creation of frameworks, resources, and simple governance that will make GenAI integration responsible, fair, and effective.
At its heart, this is more than a technological change—it is a new way of thinking about the purpose of higher education. In a world where AI can create content easily, education’s true value is in building the human skills that still matter: critical judgement, ethical reasoning, and the ability to act well in complex, uncertain situations.
We would be happy to provide oral evidence to the committee if that would be helpful.
Key reference
Doug Specht and Gunter Saunders, being indispensable: Capabilities for a Human-AI World, The Futures Framework, HEPI Report 198, March 2026.
May 2026