The Unspoken AI Gap in Schools
Most schools now have staff and pupils using AI tools daily, but very few have decided what that means for safeguarding, data protection and teaching. Senior leaders in one UK survey reported widespread concern about plagiarism, staff training, quality control, data security and safeguarding in relation to AI use in their schools (SecEd, 2025). The Department for Education (DfE) has issued multiple guidance documents and product safety standards on generative AI in education, which would not be necessary if the system already felt confident and prepared (DfE, 2023; DfE, 2024; DfE, 2025a).
Safeguarding Risks Have Moved Faster Than Policies
Keeping Children Safe in Education 2026 names AI explicitly when it describes online risks, including harmful or inappropriate content generated or amplified by AI tools and chatbots (DfE, 2026). The guidance expects schools to protect pupils from content, contact, conduct and commercial harms, and accompanying DfE training modules show how AI can introduce each of these in new ways (DfE, 2025b).
Ofsted’s published approach to AI makes clear that inspectors will look at how leaders considered risks and governance around AI before and during implementation, not only after an incident has occurred (Ofsted, 2025). A school where pupils freely access generative tools but the safeguarding policy does not mention AI is exposed, even if nothing has gone visibly wrong yet (DfE, 2026; Ofsted, 2025).
Data Protection Exposure Is Quiet Until It Is Not
DfE guidance on generative AI in schools states that tools must comply with data protection legislation and the school’s own privacy notice, and advises that personal data should not be used in generative AI tools unless strictly necessary (DfE, 2024). Schools are instructed to seek advice from their data protection officer, check whether a tool type is approved, understand how personal data is processed and ensure that data is not reused for further training (DfE, 2024).
Local authority guidance and specialist data protection advice emphasise that personal and sensitive data must not be entered into generative AI tools, and that AI‑related processing which materially affects pupils is likely to trigger a Data Protection Impact Assessment under UK GDPR (Education Data Hub, 2024). Recent analysis of new DfE AI standards translates this into specific expectations: schools should be able to explain the logic behind any automated outcome involving AI, assess algorithmic bias, avoid using pupil work as training data without consent and ensure that contracts with AI providers do not quietly assign away intellectual property rights (Data Protection Education, 2025; Data Protection Education, 2026).
Pupils Are Not Clear Where The Lines Are
An Oxford University Press study of UK teenagers found that only 44 per cent believe using AI to complete all homework counts as cheating, and only 15 per cent feel they have had adequate guidance from school on acceptable and unacceptable AI use (Oxford University Press, 2026). Those figures describe a cohort already using AI while uncertain about where academic integrity begins and ends.
The problem is visible in higher education. A national investigation reported almost 7,000 proven cases of AI‑assisted cheating in UK universities within a single academic year, with commentators describing these as the visible tip of a larger problem (Vernon, 2025). Schools relying on bans or detection tools without clear teaching on integrity are managing the symptom rather than the pattern that starts earlier in pupils’ careers (Oxford University Press, 2026; Vernon, 2025).
Staff Know They Lack The Confidence They Need
The SecEd survey showed that school leaders’ biggest concerns about AI were plagiarism, staff training and quality control, alongside data security and safeguarding (SecEd, 2025). Leaders used risk and compliance language more than innovation language when describing AI, which is itself a signal that staff do not feel equipped to use these tools well (SecEd, 2025).
Ofsted’s early‑adopter study reported that among teachers who had not started using AI, the most common reasons were lack of understanding and specific worries about privacy, bias and safeguarding (Ofsted, 2025b). These are not teachers trying to avoid change, but teachers who can see the risks clearly without structured time or support to learn how to manage them (Ofsted, 2025b; SecEd, 2025).
Bias And Deskilling Sit Behind The Headline Risks
The National Governance Association’s AI toolkit warns that AI systems can replicate and amplify existing bias, and highlights the risk of cognitive deskilling if pupils lean on AI tools instead of developing their own thinking (National Governance Association, 2026). Updated DfE generative AI product safety standards include expectations around cognitive development, emotional and social development and mental health, and instruct developers to mitigate these harms when designing products (DfE, 2025a).
The standards expect products to avoid manipulative strategies, detect signs of distress and trigger appropriate safeguarding responses, and insist that tools do not use pupil work for commercial training without consent (DfE, 2025a). When schools adopt tools without asking detailed questions about bias, deskilling and developmental impact, they take on those risks without seeing them clearly (National Governance Association, 2026; DfE, 2025a).
Complaints Are Getting Sharper And Faster
A recent school leaders survey by Browne Jacobson reported growing pressure from policy change, financial strain and AI‑related parental complaints (Browne Jacobson, 2026). Leaders described more detailed, more assertive complaints, some clearly written with or by AI tools, arriving faster than previously and adding volume to an already intense workload (Browne Jacobson, 2026).
These complaints often focus on digital and AI issues, from homework support to filtering and monitoring, and heads are expected to respond with clear, confident explanations of school decisions (Browne Jacobson, 2026). That is difficult when AI use across the school has grown informally without a documented strategy, and when contracts with providers do not explicitly state how pupil data and intellectual property are handled (Data Protection Education, 2025; ICO, 2026).
Costs And Procurement Decisions Are Under Real Pressure
The same Browne Jacobson survey found leaders worried about long‑term financial sustainability and the impact of funding on staffing, at a time when digital and AI tools compete directly with core provision for scarce budget (Browne Jacobson, 2026). DfE generative AI product safety standards were written for suppliers, but explicitly encourage schools to use them when assessing which AI products are safe and worth buying (DfE, 2025a).
The standards cover filtering, monitoring, security, privacy, intellectual property, design, governance, cognitive development and emotional impact, and expect written answers to each area (DfE, 2025a; Data Protection Education, 2025). When procurement decisions are based on a demo and a general assurance of compliance rather than the standards checklist, schools can end up paying for tools that are unsafe, unused or misaligned with classroom need (Edves, 2025; ICO, 2026).
What Sits Underneath All Of This
Across safeguarding, data protection, academic integrity, staff confidence, bias, complaints and cost, the pattern is the same. Staff and pupils use AI tools widely, and vendors build features quickly, while schools have not yet decided who owns AI strategically, what is approved and how risk is managed.
DfE guidance already sets out the pieces that need to exist: a clear position on generative AI in education, data protection rules that cover AI, product safety standards, training modules on safe use and expectations for filtering and monitoring (DfE, 2023; DfE, 2024; DfE, 2025a; DfE, 2025b). Local guidance from data protection and governance bodies turns that into practical steps such as updating privacy notices, performing DPIAs for high‑risk AI processing and checking contracts for training‑data and intellectual‑property clauses (Education Data Hub, 2024; Data Protection Education, 2025; National Governance Association, 2026).
The schools furthest ahead did not begin with a tool. They began with one decision in a room with the right people: who owns AI in this organisation, and what do we expect of it in our school.
References
Browne Jacobson (2026) School Leaders Survey summer 2026 results: Policy, funding and AI. Available at: https://www.brownejacobson.com/insights/school-leaders-navigate-send-reform-financial-pressures-and-the-rise-of-ai-generated-complaints (Accessed: 9 September 2026).
Data Protection Education (2025) New AI guidance from the DfE. Available at: https://dataprotection.education/best-practice-update/new-ai-guidance-from-the-dfe (Accessed: 9 September 2026).
Data Protection Education (2026) New DfE AI Standards – best practice update. Available at: https://dataprotection.education/best-practice-update/new-dfe-ai-guidance (Accessed: 9 September 2026).
Department for Education (DfE) (2023) Generative artificial intelligence (AI) in education. Available at: https://www.gov.uk/government/publications/generative-artificial-intelligence-in-education (Accessed: 9 September 2026).
Department for Education (DfE) (2024) Generative artificial intelligence (AI) and data protection in schools. Available at: https://www.gov.uk/guidance/data-protection-in-schools/generative-artificial-intelligence-ai-and-data-protection-in-schools (Accessed: 9 September 2026).
Department for Education (DfE) (2025a) Generative AI: product safety standards. Available at: https://www.gov.uk/government/publications/generative-ai-product-safety-standards (Accessed: 9 September 2026).
Department for Education (DfE) (2025b) Safe use of generative AI in education: module 3. Available at: https://www.gov.uk/government/publications/safe-use-of-generative-ai-in-education-module-3 (Accessed: 9 September 2026).
Department for Education (DfE) (2026) Keeping children safe in education 2026. Available at: https://assets.publishing.service.gov.uk/media/6a9081309a177a1decf97b00/Keeping_children_safe_in_education_2026.pdf (Accessed: 9 September 2026).
Edves (2025) DfE generative AI standards. Available at: https://edves.uk/compliance/dfe-generative-ai-standards/ (Accessed: 9 September 2026).
Education Data Hub (2024) AI Guidance for Schools. Available at: https://educationdatahub.org.uk/wp-content/uploads/AI-Guidance-for-Schools.pdf (Accessed: 9 September 2026).
Information Commissioner’s Office (ICO) (2026) EdTech audit finds widespread compliance failures in children’s services. Available at: https://www.info-gov.uk/data-protection/data-protection-news/ico-edtech-audit-finds-widespread-compliance-failures-in-childrens (Accessed: 9 September 2026).
National Governance Association (NGA) (2026) NGA AI toolkit. Available at: https://www.nga.org.uk/media/5swprdaw/nga-ai-toolkit-100726.pdf (Accessed: 9 September 2026).
Ofsted (2025) How Ofsted looks at AI during inspection and regulation. Available at: https://www.gov.uk/government/publications/ofsteds-approach-to-ai/how-ofsted-looks-at-ai-during-inspection-and-regulation (Accessed: 9 September 2026).
Ofsted (2025b) ‘The biggest risk is doing nothing’: insights from early adopters of artificial intelligence in schools and further education colleges. Available at: https://www.gov.uk/government/publications/ai-in-schools-and-further-education-findings-from-early-adopters (Accessed: 9 September 2026).
Oxford University Press (2026) UK teens and AI homework: attitudes to cheating and integrity. Summary reported in: AI Weekly Alert, 8 June 2026. Available at: https://aiweekly.co/alerts/oup-study-uk-teens-split-on-ai-homework-cheating (Accessed: 9 September 2026).
SecEd (2025) ‘Schools in AI data protection and compliance warning’, SecEd, 25 February. Available at: https://www.sec-ed.co.uk/content/news/schools-ai-data-protection-and-compliance-warning (Accessed: 9 September 2026).
Vernon, A. (2025) ‘Thousands of UK university students caught cheating using AI’, The Guardian, 14 June. Available at: https://www.theguardian.com/education/2025/jun/15/thousands-of-uk-university-students-caught-cheating-using-ai-artificial-intelligence (Accessed: 9 September 2026).