.jpg)
As an emerging body of literature suggests, artificial intelligence is radically revolutionising childhood and adolescence, causing radical alterations and disruptions to the way that young people socialise, think, and interact with the world around them. As AI generative tools seek to radically conquer pedagogical interactions and environments with pristine algorithmic efficiency, I’ll assume a universal consensus; the speed of such developments should, perhaps, surprise just a few.
In fact, the Department for Education (DfE) embraces these developments openly. In a recent press release dated to the end of January, the DfE has announced radical reforms to implement ‘safe AI tutoring tools co-created with teachers’ to support classroom learning across England, but with a particular emphasis on supporting children from underprivileged backgrounds, unable to access private tutoring services due to financial restraint. As the educational attainment gap widens between the underprivileged and their wealthier peers, the DfE hopes to alleviate these disparities by integrating personalised AI tutors into classrooms across schools in the UK, offering personalised one-to-one support that is ‘often only available to the privileged few’.
The DfE states that too many disadvantaged children appear to lag significantly behind their peers ‘with just one in four achieving a pass in English and Maths at GCSE at grade 5 or above, as compared to over half of their peers’. The DfE cites a research study from the Education Endowment Foundation which demonstrates that personalised tutoring support proves to be mightily effective intervention for improving pupil outcomes. Existing findings indicate that tutoring can accelerate a pupil’s learning by around 5 months.
The EEF stresses, however, that tutoring is more likely to have an impact as a supplementary learning resource used to support existing classroom teaching and lessons. The intention must not be to replace whole classroom teaching with personalised tuition but rather supplement the existing pedagogical process with additional benefits that one-to-one or even small group teaching support can provide. Create an existing space where teaching can be tailored to directly address individual needs and misconceptions, interaction is personalised, and progress is a mightily tangible and visible outcome.
Overall, there appears to be very little ambiguity with how these tools are expected to operate in classrooms. Bridget Phillipson, Education Secretary, emphasises that such tools must ‘never replace the human connection that only great teachers can provide.’ Sarah Hannafin, head of policy at the school leaders’ union NAHT, stresses in her response to the DfE press release that they must ‘never replace face-to-face interaction.’
The press release from the DfE is certainly optimistic about the impact of AI tutoring and, perhaps, the optimism appears to have some rationale behind it. However, keep in mind, this process is still in its infancy. There remains substantial ambiguity in regards to the exact ‘AI tools’ or ‘digital companions’ that the DfE envision in their usage of artificial intelligence. Although the tools are to be made available to schools by the end of 2027, ‘teacher-led co-creation of AI tutoring tools’ with the expertise of ‘leading tech companies’ and ‘AI labs’ will only begin in the summer term of 2026. We will perhaps have to wait until much later in the year for an opportunity to rigorously evaluate the ‘tools’ once they come to life.
However, more concerning than the supposed ‘prematurity’ of the optimism with little idea as to how these ‘AI tools’ will deliver at present, does there appear to be a disconnect between aspirational policy ideas and on-the-ground practical constraints that may limit the efficacy of these tools once they are implemented?
Recent national and local authority budget cuts to education may undermine the practical efficacy of AI in education. Schools are expressing their profound worries as they are having to cut back on teaching, teaching assistants, and support staff. As The Sutton Trust reports, 51% of secondary senior staff have reported making cuts in teaching staff, with cuts at primary schools being the highest the trust have seen since the inauguration of this poll in 2017. Furthermore, there has been a substantial increase in secondary and primary school leaders reporting cuts to IT equipment.
It appears the reforms assume the schools have the personnel and infrastructure to integrate AI responsibly, but in this climate, systemic underfunding may jeopardise its efficacy with resource-dependent policy being introduced into resource-constrained environments. If schools end up adopting AI tutoring tools out of ‘regulatory’ or ‘national policy pressure’ to reform curricula, could certain schools become inclined towards total dependency on AI to ‘plug’ the teaching staff gap? Without parallel investment to maintain the necessary IT infrastructure and staffing numbers to provide AI support safely and equally to all students across the country, the optimism may very well be short-lived.
Similar concerns are echoed by Daniel Kebede, general secretary of the National Education Union (NEU). Kebede and the NEU have stressed concerns that AI teaching tools could lead to some in the profession losing work, such as teaching assistants. While they welcomed developments that could ‘free up an educator’s time’ more effectively, such as reducing administrative and bureaucratic workload, Kebede stresses ‘education and learning is ultimately a relational and social experience’. They would resist any such ‘direction of travel that [would] seek to de-professionalise, deskill or replace teaching assistants.’ If AI can be trained to handle much of the personalised interaction and support that pupils receive and greatly benefit from, with total algorithmic efficiency, precision, and 24/7 availability, then there’s a very real risk that many of the almost 300,000 full-time teaching assistants employed in England could be out of work very soon.
While the DfE have hit back at Kebede and the NEU, stressing that all such policies have been thoroughly developed in consultation with teachers, such concerns have not emerged in isolation. When schools across the country continue to face substantial fiscal crises and funding shortages, some teachers are questioning why AI provision is even being prioritised at a time when schools are ‘struggling to afford even the basics.’ The risk of compulsory redundancies, attributed to severe fiscal crises in schools, have forced teachers across the country to strike over the years, while also demanding pay rises as the profession risks losing hundreds of teaching staff each year to poor, inadequate pay. Yet, do the optics of modernisation just matter more than ensuring classrooms are adequately resourced at the baseline and teachers are adequately reimbursed?
Now, I believe we’re also missing the input of an audience grossly underrepresented in decision-making. While the press release makes little mention of consulting pupils in the development of these AI tools, I should hope there is intention to consult those who will ultimately be engaging and interacting with the tools once they come to life. What do the children say?
Overall, children of all Key Stages recognise the potential of AI and how it could enhance learning experiences. A research report from the DfE and DSIT (Department for Science, Innovation and Technology) in 2024 provides detailed findings exploring parent and pupil attitudes in regards to the use of AI in education. Pupils found AI tools to be ‘exciting and engaging’ and expressed interest in how these tools could help bring pupil-generated ideas to life, bouncing off of pupil-generated ideas to develop creative essay or story plans which could inspire pupils further. Additionally, pupils felt that AI could provide immediate and tailored support to pupils’ needs and could help with providing personalised support in a topic or subject they were struggling with. For example, pupils felt they could benefit from AI-tutor generated ‘extension activities’, ‘summary sheets’ or mini quizzes to consolidate a pupil’s learning after a lesson. As one Year 6 student expressed in a recent 2025 Student Engagement Report commissioned by the DfE, ‘AI can’t get mad at you’, reflecting perhaps, a safe space for pupil trial, error, and exploration due to AI’s inability to express human emotions of anger and impatience.
However, pupils have also been quick to express their concerns with the use of AI in education, and the post-COVID generation of students have been keen to maximise face-to-face learning experiences with their teachers after years of confinement to virtual classrooms. Isolation from peers and teachers has resulted in many young people experiencing a rapid deterioration in their mental health. As one GCSE pupil said, ‘I missed the social interaction of being in schools, during the lockdowns implemented in response to Covid-19.’ 80% of pupils in the Student Engagement Report agreed that AI should never replace ‘face-to-face teacher interaction,’ with students preferring human teachers. One Year 12 student expressed that the future of education with AI seemed ‘dystopian’ for AI was ‘dehumanising’ in taking away the student-teacher relationship. Of course, where younger children may not be as knowledgeable about AI, it is sensible that parents are consulted extensively as these tools begin come to life. Parents may have rightful concerns about how they wish for their children to interact with AI and deserve to have a space at the table as these tools are developed.
Overall, it’s an exciting development, but many are right to express major concerns. I should hope the DfE will also consult with pupils as they develop these tools as young people also deserve to be meaningfully involved in decision-making relating to their learning. Their voices and concerns are equally valuable, and pupils can reveal whether these tools feel supportive, isolating, empowering, or anxiety-inducing. Yet, as teachers have pointed out, are policymakers ‘jumping the gun’ when classrooms remain severely under-resourced at the very baseline?
.svg)
.svg)
.svg)
