PRINCIPLES OF ETHICAL AI USE
For internal quality assurance, define universally applicable PRINCIPLES OF AI USE, through an assessment of the risks and opportunities of AI across your school, academy or trust.
The principles set out below are also available as a downloadable exemplar set of ‘Ethical Principles’: See ‘Step 2’ of the ‘Three Steps Model‘ section of this website.
Wisdom, Safeguarding, Data & Privacy…
Protect Wisdom, Truth and Human Dignity: AI replicates the best and the worst
A Christian ethical approach to AI begins with the belief that every person is made in the image of God. AI must never reduce pupils to data points or replace relational work. Wisdom requires discernment; truthfulness requires transparency; and dignity requires boundaries.
Key Principle: AI must serve people, protect the vulnerable, and uphold the school’s Christian vision.
Practical Action: Ensure that the AI Vision, AI Assurances and later consequential implementation models place a strong emphasis on dignity and respect.
Understand New Safeguarding Risks: AI as an Online Risk
AI introduces new safeguarding risks: harmful content, deepfakes, grooming, misinformation, personalised emotional manipulation, and data misuse. These risks must be embedded in safeguarding systems, not treated as technical issues.
Key Principle: AI incidents are safeguarding incidents.
Practical Action: Staff must be trained, DSLs must lead, and pupils must be protected: Integrate AI into all existing safeguarding policies, processes, practices and training.
Filter and Monitor Well: What “Appropriate” Means
Filtering prevents access to harmful content; monitoring detects concerning behaviour. AI complicates both because harmful content can now be generated internally.
Key Principle: Leaders must ensure systems are safe, proportionate, transparent, and reviewed annually. DSLs must understand how alerts work. Over‑blocking harms learning; under‑blocking harms pupils.
Practical Action: Review thresholds, test robustness, and publish human as well as technical expectations.
Protect Children’s Privacy: by design (use the Children’s Code)
The Information Commissioner’s Office (ICO) Children’s Code requires high privacy by default, transparency, minimisation, and avoidance of manipulative design. AI tools must be configured with child safety at their core. Suppliers must meet Department for Education (DfE) product safety expectations.
Key Principle: AI must protect children’s rights, not exploit them.
Practical Action: Ensure that all AI used at minimum meets the ICO and DfE requirements.
Protect Data Rigorously: Lawful, Fair, Transparent AI Use
Artificial intelligence brings extraordinary capability into schools, but it also introduces significant data protection challenges — especially when the data involved belongs to children. UK GDPR requires all processing to be lawful, fair, transparent, and accountable. AI complicates these duties because it can process data in opaque ways, generate inferences that are difficult to explain, or use personal information in ways that are not immediately visible. AI raises significant data protection challenges, Ethical AI use therefore requires strong data governance, clear boundaries, and careful oversight.
Key Principle: Privacy is a matter of dignity and trust.
Practical Action: Ensure AI is an integral part of GDPR policy and compliance procedures.
Data Risk Assessment: Using A Risk Register (e.g. DPIA)
AI introduces safeguarding, privacy, bias, and operational risks. Risk registers help leaders assess these risks before adoption. Mitigations must be clear, realistic, and documented. High-risk tools require heightened scrutiny.
A Data Protection Impact Assessment (DPIA) is required whenever processing personal data is likely to result in a high risk to individuals. AI tools frequently fall into this category because they may analyse, store, or generate inferences about personal information. A DPIA Trigger Checklist helps leaders decide when a DPIA is necessary and ensures compliance with UK GDPR. In a church school, the checklist also supports ethical reflection on dignity, fairness, and stewardship.
Key Principle: Effective risk assessment transforms caution into wisdom.
Practical Action: Ensure that AI is a named element within any oversight strategic risk register, and has its own discrete AI risk register.
Teaching, Curriculum, Inclusion, Learning, Assessment, Truth Skills, Pastoral Care…
Plan Staff Use: Teaching, Planning, Admin and Professional Judgement
AI can act as a virtual tutor, assistant, or practice partner — but never as a teacher in the full relational, moral, and spiritual sense. Teaching is incarnational, embodied, and accountable. AI can support learning, but it must never replace presence. AI can reduce workload — but it must never replace human time invested in wise professional judgement.
Staff must verify outputs, protect data, and maintain self-driven standards. AI is a tool, not a replacement for humans, a shortcut to maintaining standards, or a line manager.
Key Principle: Professional identity is integral to a sense of vocation and must not be artificially replaced.
Practical Action: Assess where AI needs to feature in habitual training (such as safeguarding) and ensure that all AI training and development goes beyond technical skills to address the ethical dimension, as well as the institution’s assurances and implementation models.
Integrate into the Curriculum: cross-subject AI literacy
AI literacy must encompass critical evaluation, ethical awareness, safe practice, and theological reflection. Pupils must learn to question AI, recognise bias, and seek the truth.
Key Principle: Wise discernment is the heart of a Christian digital curriculum.
Practical Action: Evaluate AI’s place in all curriculum plans, linking it to your AI assurance statement (or equivalent), ethical use implementation approaches, and any learner/pupil/student charter.
Focus on Inclusion and SEND: Opportunities and risks
AI can support accessibility — but it can also introduce new vulnerabilities. Pupils with SEND may be more trusting of AI or more vulnerable to harm. AI must enhance inclusion, not replace human support.
Key Principle: Inclusion is a theological commitment, not a technical feature.
Practical Action: Ensure that AI is a named part of your policies and statements related to all provision for SEND.
Control Bias, Discrimination and Justice: Equity vs equality
AI reflects human fallibility through its training data. It can disadvantage groups, reinforce stereotypes, or misinterpret language. Church schools must test tools, mitigate bias, and teach all staff and pupils to recognise injustice. AI seeks to please and can be complicit in feeding user-led confirmation bias.
Key Principle: Addressing bias is an act of courageous advocacy: It is the responsibility of all staff to monitor against prejudice.
Practical Action: Ensure that AI is a named part of your policies and statements related to the delivery of the Equalities Act, and include AI in any equality and inclusion training.
Synthetic Learning: The risks of simulated progression
AI can support learning, but it can also undermine the learning process. In particular, the capacity of generative AI to propose ideas and articulate them clearly and proficiently provides a powerful but potentially corrupting substitute for thinking. By enabling learners to temporarily bypass important personal and cognitive stepping-stones, it fosters false knowledge, skills, and understanding.
Such vicarious or disconnected personal progression can be damaging if left unaddressed, leaving gaps in an individual’s cognitive and physical learning journey. Shortcuts can be attractive to learners, but in a carefully sequenced curriculum, they disrupt or even render the learning experience neutral.
Key Principle: A learner’s personal and intellectual development journey cannot be outsourced; AI must not be allowed to falsely simulate the acquisition of knowledge, skills, and understanding.
Practical Action: Ensure that any teaching and learning activity is covered by your institutional policy on acceptable AI use at a pedagogic – not just technical – level, particularly ensuring that clear expectations and boundaries govern the use of tools that undertake artificial thinking for a learner.
Assessment Reliability: Helping vs cheating
AI can undermine honesty and the validity of assessments. Schools must set clear boundaries, design AI-resilient assessments, and teach pupils why integrity matters. Defining what constitutes effective and appropriate use of AI across both summative and formative assessments is critical.
Key Principle: Truthfulness is a Christian virtue.
Practical Action: Ensure that any assessment activity is covered by your institutional policy on AI, and that clear expectations and boundaries are set on its use, including employing tools to screen for AI use where relevant. Use your AI assurance statement and ethical framework to benchmark suitability.
Address Specialist Risks: Deepfakes, misinformation and online harms
Deepfakes threaten truth, dignity, and trust. Pupils must learn to recognise manipulation. Schools must respond swiftly, pastorally, and transparently.
Key Principle: Truthfulness is a non-negotiable spiritual and moral commitment.
Practical Action: Ensure that AI is a core element of the curriculum, providing adults and learners with the knowledge, skills and understanding to critically evaluate and recognise AI-generated and/or fake content.
Nurture Pastoral Care: Boundaries and Relational Integrity
AI must never be used for pastoral guidance, safeguarding notes, or emotional support. Pastoral care is relational, embodied, and accountable. AI cannot replace presence, empathy, or spiritual discernment.
Key Principle: Pastoral care must remain human.
Practical Action: Ensure that AI assurances and your ethical framework impose strict controls and boundaries on the use of artificial sources of pastoral support.
Governance, Communication, Stakeholder Engagement and Record Keeping…
Have Strong Stewardship: Exercise an Assurance Governance approach
Governors must ensure that AI use is aligned with the institution’s vision and is safe, fair, and well-governed. They must ask about safeguarding, data protection, bias, procurement, and impact. They do not need technical expertise — only wisdom, curiosity, and courage.
Key Principle: Strong stewardship is a core Christian principle exercised through governance that protects dignity and ensures accountability.
Practical Action: Insist that governors and/or trustees co-create the AI Vision, and approve the AI assurance statement and ethical framework. Make AI awareness a core part of governor/trustee development and training, particularly those with nominated roles such as safeguarding and SEND, etc.
Deploy a Values-Driven AI Acceptable Use Policy: Distinctive, clear and supportive
An AI policy must be theological, ethical, safeguarding‑led, and operationally clear. It must set boundaries, define responsibilities, and ensure transparency.
Key Principle: Policy turns values into practice. Authentic beliefs have a lived impact.
Practical Action: Create an AI Acceptable Use Policy that aligns with your AI Assurance statements and ethical framework, and review it annually.
Engage Pupils & Students in Ethical Use: Change cultures
Along with the curriculum, pupil/student charters can promote safety, honesty, respect, privacy, discernment, and Christian values. It helps pupils take equal ownership of their role as wise, responsible digital citizens.
Key Principle: Charters shape culture and a sense of belonging.
Practical Action: Co-create an AI Learner/pupil/student Charter aligned with your AI Assurance statements and ethical framework. Use it to inform a learner-voice approach to monitoring risks, concerns and user experiences.
Work with Stakeholders: Parents & carers, and the wider community
Parents need clarity, support, and partnership. Schools must educate families about AI risks, boundaries, and opportunities. Community resilience grows through shared understanding.
Key Principle: Partnership strengthens flourishing.
Practical Action: Provide training and awareness-raising sessions for parents and carers, using a parent and student agreement on the appropriate use of AI, aligned with your AI Assurance statements and ethical framework.
Communicate Well: Trust,Tone and Transparency
AI can enhance clarity and accessibility — but it can also misjudge tone or undermine authenticity. Staff must disclose AI use, verify accuracy, and avoid using AI for pastoral or disciplinary messages.
Key Principle: Communication must remain relational and truthful.
Practical Action: Ensure that AI policy and staff handbooks include clear guidance on the appropriate use of AI to formulate, edit or control internal and external communications.
Record AI Incidents: Responding, recording, learning
AI incidents must be logged, investigated, and used for learning. They require safeguarding responses, root‑cause analysis, and transparent communication. This requires a safe-space culture of professional improvement, not an anxious subservience to compliance driven by fear. Mistakes happen when change is rapid; the hope is to minimise their impact and recurrence.
Key Principle: Incidents may have implications, but they are also learning opportunities. Christian integrity requires us to be honest with ourselves when things go wrong.
Practical Action: Ensure that AI is fully integrated into your safeguarding, prevention, and reporting processes.
Training, KCSIE, Inspection, Procurement, Cybersecurity…
Prioritise Staff Training: A KCSIE-linked approach
Training must cover safeguarding, safe use, curriculum integration, professional judgement, pastoral boundaries, and theological reflection. AI literacy is essential for all staff, which includes how to use evaluative models to assess impact and undertake responsible experimentation.
Key Principle: Training builds confidence and consistency. We owe that to ourselves and others.
Practical Action: Introduce initial AI induction training, followed by a cycle of annual update training as a stand-alone and/or integrated component of all staff training – as would be the case for safeguarding.
Collate Safeguarding Evidence (KCSIE): A litmus test of care
AI risks must be embedded within safeguarding systems. Evidence includes training, incident logs, filtering reviews, DSL oversight, and pupil education.
Key Principle: Safeguarding is the foundation of ethical AI.
Practical Action: Ensure that AI is an integral component of all safeguarding policies, processes and culture.
Consider Ofsted: Impact-focused scrutiny
Ofsted focuses on impact, not innovation. Inspectors ask how AI impacts learning, safety, assessment, and inclusion – they are not prescriptive on use or platform. Leaders must demonstrate informed, responsible decisions. Integrate AI into provision by adopting the same ‘thinking approach’ used across all areas of school, academy, or trust life.
Key Principle: AI must support high‑quality learning, not define it. Ofsted evaluates outputs, not inputs.
Practical Action: Use the standard Ofsted toolkit and judgement statements to assess how and where AI might be considered a constituent part of Ofsted’s focus, with particular attention to safeguarding, assessment and its positive contribution, or otherwise, to learner outcomes.
Use for SIAMS Evidence: Vision-led AI
AI use can strongly evidence a living Christian vision — when it is intentional, ethical, and relational. SIAMS looks for theological language, dignity, justice, and formation.
Key Principle: How AI is implemented and used is an important exemplar of this vision in action.
Practical Action: Use the SIAMS IQs as the basis to assess how and where AI might be considered a constituent part of distinctly Christian provision, with particular attention to setting the overall vision, safeguarding, and its positive contribution, or otherwise, to achievement in all areas.
Procure Carefully: Enacting Due Diligence
AI procurement must be purpose-led, risk-based, and ethically grounded. DSLs and DPOs must be engaged. Tools must be tested, documented, and reviewed. Suppliers must demonstrate transparency and ensure safety.
Key Principle: Procurement is a moral task, not a technical one.
Practical Action: Ensure that all AI procurement follows statutory requirements, incorporating AI into any current procurement and commissioning policies and procedures.
Deploy Tight Cybersecurity: Protect systems and identities
Cybersecurity is about safeguarding. Schools must protect data, train staff, assess suppliers, and maintain system integrity. Stewardship requires vigilance.
Key Principle: Digital safety protects the vulnerable, all of whom are our neighbours.
Practical Action: Ensure that habitual training and development is undertaken in the safe use of AI, and the potential for AI to be used to mislead, corrupt and appropriate.
Need more help?
Please use the contact page if you would like to discuss any of the below…
1:1 Support or Staff Workshops
Understand the core principles of responsible AI use in a focused 1-hour session.
Conference / Group Speaking
Deploy a dynamic speaker with recent school and trust leadership experience.
Access the Resource Library
Download expert guides and materials to deepen your knowledge.