Written evidence by Dr Kwok Tong Soo (AIE0015)
Education Committee
The use of Artificial Intelligence and EdTech in Education
Dr Kwok Tong Soo is a Senior Lecturer in Economics at Lancaster University, and currently Director of Teaching in the Department of Economics. He has developed the Department’s AI Strategy and EDI Strategy, and is a member of the University’s Academic Integrity Network. He has research interests in the Economics of Higher Education, alongside other research interests in International Trade and Regional and Urban Economics. This evidence is based on his academic knowledge and understanding of the issue and is given in a personal capacity[1].
Where specific products or services are referenced, this is for illustrative purposes only and does not constitute endorsement.
Since the release of ChatGPT in November 2022, adoption of generative AI has been extremely rapid. A recent survey indicates that over 90 percent of UK university students use AI in some way[2]. However, the same survey indicates that universities’ support for AI skills lags behind. Overcoming this lag will require significant investment, not only in IT equipment, but especially in governance structures, training, guidance and AI literacy of teaching, research and support staff, at all levels of education. The debate over whether AI should be used in education should be progressed to how AI should be used at each level of education.
1.1. The key opportunities shaping the development and adoption of EdTech and AI across the education system
The key opportunity is that AI has the potential to enable more personalised education at all levels, enhancing learning and comprehension, preparing students for assessment, and developing their research and information skills. It may also help teachers with useful tools to help them be more efficient, in lesson planning and preparation, assessment and feedback, administrative and professional work, and whole-school and strategic uses. The use of AI may after the initial adjustment period help reduce the overload faced by teachers (survey data indicates the average school teacher works over 50 hours in a typical week[3]).
1.2. The most significant challenges and risks currently shaping the use of EdTech and AI in the education system
There are many challenges. AI is rapidly developing, so everyone involved has to invest significant time and resources in order to keep up. This calls for educational providers to devote more resources to staff training in the uses of AI, and provide appropriate institutional support. One key challenge in doing this, is the development of clear governance structures that balance ISS (Information Systems Services) teams who bring technical expertise but may prioritise risk avoidance over educational value, and academic users who understand pedagogical need but may lack technical literacy. Without clear governance structures, decisions may get made by default rather than by design, and the people closest to teaching and learning may have limited voice in decisions that profoundly affect them.
Equity and access is also a key challenge. If AI genuinely confers educational advantage, then unequal access risks widening existing attainment gaps, if students from wealthier backgrounds, in better-resourced schools and universities, with more tech-literate parents and teachers, get more and better AI support than their less advantaged peers.
1.3. The current state of digital infrastructure across schools, colleges, universities and other education settings.
The digital infrastructure, broadly defined, includes not only the hardware required for these AI systems, but also the governance systems, and the training and literacy of both staff and students; in other words, all that is required to make effective use of AI systems[4]. Defined as such, many educational institutions at all levels in the UK do not currently have the digital infrastructure to fully capitalise on using AI in education. As discussed above, a big gap is that between student use of AI, and institutional guidance and training on the use of AI. Another big gap is that of institutional governance structures, and wider regulatory guidance and national coordination on the use of AI in education.
2.1. The impact AI is having on teaching practices.
The current use of AI in teaching mainly revolves around the use of AI to help develop teaching content. This may include classroom materials and quizzes. New specialist AI tools can also help with marking and administrative tasks, freeing up teachers’ time for developing and delivering teaching content. For most part, we have only started to explore the possible uses of AI in education.
2.2. The confidence of teachers and other education practitioners in using AI within their practice.
This varies widely across teachers. CPD infrastructure to systematically address this variation is currently underdeveloped. Individual enthusiasts do their own learning, while colleagues with less initiative or time are left behind.
2.3. Whether teachers, lecturers and other practitioners currently have enough support and training on how to use AI effectively, including initial teacher training and CPD.
For most part, no. Much more investment is needed to support and train practitioners on how to use AI effectively. This investment need not necessarily be financial: for example, the AI firm Anthropic offers free online courses in a variety of AI skills[5]. What is really required is time set aside for the express purpose of staff learning how to use AI effectively, along with appropriate guidance and governance structures.
2.4. How the growing availability of EdTech and AI in education settings is affecting the reliability and validity of traditional assessment methods – including coursework completed outside the classroom.
This is a key issue around the integrity of academic qualifications: How do we know if the assessment has been completed by the learner, or by an AI tool? As AI tools become integrated into education, this question becomes less meaningful, as successful completion of the assessment may require appropriate use of AI tools. The assessment tasks themselves must change, probably to include greater emphasis on the process (how the submission is put together) rather than just the final product. In the longer term this may involve the integration of AI in assessment. The focus becomes why the learner has used particular tools or prompts, how they have changed their prompts in response to AI feedback, and learner self-reflection on the process[6]. This reframes the assessment strategy as evidencing students’ AI fluency.
2.5. How the national curriculum may need to adapt in light of the increasing use of AI, and how the skills that students may need to succeed in the workplace might change. Did the Curriculum and Assessment Review adequately anticipate the adoption of AI or is there a need for further focused work in this area?
In my view the Curriculum and Assessment Review[7] did not adequately anticipate the adoption of AI, and further focused work is required in this area. The national curriculum needs to do more in terms of providing students with digital literacy skills. The time may have come to consider making Computing a compulsory GCSE subject. This will require a large investment in computing facilities and staff in schools, but it represents an investment in future capacity building, and will provide the country with a workforce that is well-placed to lead the AI revolution, in line with the UK government’s AI Opportunities Action Plan[8]. Similarly, many universities are starting to introduce modules in AI literacy and fluency.
2.6. The continuing professional development (CPD) needs that arise from the increased adoption of AI in teaching.
As mentioned above, increased adoption of AI in teaching at all levels will require a large increase in investment in staff training in the form of CPD. Similarly to the investment in computing facilities and staff in schools, increased CPD should be viewed as investment in the capability of our educational institutions to deliver a workforce that can lead the AI revolution.
3.1. The ways in which AI is reshaping learning and education for learners of all ages.
3.2. How AI-enabled tools may influence the development of essential skills such as oracy, problem solving and creativity, and how these tools could be used to support emerging methods of documenting and recognising competencies.
There are many ways in which AI can enhance the development of essential skills. For example, most AI tools can be tuned to a “Socratic” mode, where they are programmed to respond to prompts with a question or a hint, rather than simply providing the answers. AI is also effective at brainstorming ideas, and can often throw up suggestions that the user may not have considered. It could even be used to enhance critical thinking, for example if they are set up to provide a “Red Team” to challenge pre-conceived views[9].
More generally, the ability of AI tools to respond to individual learners means that they can adapt to the learner. This can improve student engagement and enhance academic performance, and the use of AI tools can significantly reduce the cost of implementing an adaptive learning system[10]. In addition to Socratic questioning, such adaptive learning systems can take the form of personalised explanations, summarising complex material, language learning, accessibility tools, practice and self-testing, feedback on writing, exam technique coaching, structured research assistance, and argument mapping. For example, AI-enhanced language learning software such as Duolingo have been effective in improving the learning of languages, through its adaptive learning system.
4.1. The evolving nature of digital risks, including algorithmic amplification, addictive design features and the impact of rapidly developing technologies such as generative AI, and the safeguarding implications that arise from them.
This is a serious concern. Safeguards should be put in place to protect young people from the digital risks from AI. These might include age-appropriate AI versions and content filtering, usage limits, AI transparency, parental or school controls, and enhanced AI literacy education. A recent trend is the concept of a “Walled Garden” AI, which are closed, secure AI systems which provide a layered safeguarding system for children. Examples include SchoolAI, TeachBetter, Khanmigo, and TeachMateAI. Such walled gardens may be a way to get educators especially in schools to embrace the use of AI[11]. In addition, there are many ways of using AI in educational settings; it may be time to advance the debate beyond whether AI should be used in education, to how it should be used, at each level of education. For example, adaptive learning systems within walled gardens can be used even from a young age, while the introduction of open-ended AI tools such as chatbots may be best deferred until learners have developed formal operational thinking, especially the metacognitive capacity to critically evaluate AI-generated content rather than accepting it uncritically. Chatbot access in educational contexts should be introduced in conjunction with AI literacy education and complementary safeguarding measures, providing learners with both the skills and the protections necessary for responsible use.
5.1. The UK government should take steps to expand the use of EdTech and AI in education at all levels, to develop a workforce which is able to lead the AI revolution.
5.2. This will require significant investment not only in physical infrastructure and ISS personnel, but also in general staff training and CPD.
5.3. The development of robust AI governance structures in educational institutions is a pressing issue, requiring closer collaboration between ISS teams and end-users.
5.4. The debate over whether or not to use AI in education should be moved on, to how AI should be used, at each level of education.
Anthropic (2025), Anthropic courses. Available at: https://anthropic.skilljar.com/ (Accessed: 13 March 2026).
Department for Education (2025), Curriculum and Assessment Review: Final Report. Available at: https://www.gov.uk/government/publications/curriculum-and-assessment-review-final-report (Accessed: 16 March 2026).
Department for Science, Innovation and Technology (2025), AI Opportunities Action Plan, Independent report. Available at: https://www.gov.uk/government/publications/ai-opportunities-action-plan/ai-opportunities-action-plan (Accessed: 13 March 2026).
Forsa, B. (2023), How do we get educators to embrace AI? Start by building a walled garden, Forbes Technology Council, Council Post. Available at: https://www.forbes.com/councils/forbestechcouncil/2023/11/03/how-do-we-get-educators-to-embrace-ai-start-by-building-a-walled-garden/ (Accessed: 13 March 2026).
House of Commons Library (2026), Teacher Recruitment and Retention in England, Research Briefing. Available at: https://commonslibrary.parliament.uk/research-briefings/cbp-7222 (Accessed: 12 March 2026).
JISC (2023), Digital Strategy Toolkit. Available at https://www.jisc.ac.uk/guides/digital-strategy-toolkit (Accessed: 13 March 2026).
McNulty, Niall (2023), Assessment for Learning: Effective Strategies and Techniques. Available at: https://www.niallmcnulty.com/2023/04/assessment-for-learning-effective-strategies-and-techniques/ (Accessed: 12 March 2026).
Ministry of Defence (2021), Red Teaming Handbook, 3rd Edition. Available at https://assets.publishing.service.gov.uk/media/61702155e90e07197867eb93/20210625-Red_Teaming_Handbook.pdf (Accessed: 12 March 2026).
Mirari, K. (2022). The Effectiveness of Adaptive Learning Systems in Personalized Education. Journal of Education Review Provision, 2(3), 107-115. Available at https://doi.org/10.55885/jerp.v2i3.194 (Accessed: 12 March 2026).
Stephenson, R. and Armstrong, C. (2026), Student Generative AI Survey 2026, HEPI Report No. 199, Higher Education Policy Institute. Available at https://www.hepi.ac.uk/wp-content/uploads/2026/03/HEPI-Report-199-Gen-AI-Survey-2026.pdf (Accessed: 13 March 2026).
May 2026
6
[1] Lancaster University webpage for Dr Kwok Tong Soo.
[2] Stephenson and Armstrong (2026).
[3] House of Commons Library (2026).
[4] JISC (2023).
[5] Anthropic (2025).
[6] McNulty (2023).
[7] Department for Education (2025).
[8] Department for Science, Innovation and Technology (2025).
[9] Ministry of Defence (2021).
[10] Mirari (2022).
[11] Forsa (2023).