Alex Dave, safeguarding lead at edtech charity LGfL – The National Grid for Learning, shares practical guidance to help schools develop a clear, effective AI policy that strengthens safeguarding and supports staff to help ensure AI is used safely across the school community

Artificial intelligence may already be embedded in most schools, whether formally adopted or not. It can appear in search engines, MIS platforms, classroom tools and communication systems. Alongside this, generative AI (Gen-AI) is increasingly being used by pupils and staff which could, if used informally or without clear guidance and oversight, lead to safeguarding and data protection issues.

For most school leaders, the question is no longer whether AI will be used, but how it’s managed.

Why does having a policy on AI matter?

Safeguarding is a primary concern for all schools. Pupils may already be engaging with Gen-AI tools outside of school, at increasingly young ages. Without guidance, they may be exposed to misinformation, inappropriate content, data misuse and unsafe interactions.

Gen-AI also presents opportunity. It can reduce workload, support accessibility, and enhance teaching and learning. However, these benefits tend to only materialise when its use is structured and intentional.

Without a clear policy, practice could become inconsistent. Some staff may interpret its use differently, pupils may become over reliant on AI, face safety, ethical and data protection risks. Expectations around assessment and behaviour may also become unclear which could lead to weakened safeguarding, academic integrity and staff confidence.

Scope and intent

The scope should apply to all staff, pupils, governors, trustees and any relevant external users. It should cover all uses of AI, including teaching, learning, assessment, administration and communication, both on and off site when linked to school activity.

It should also recognise that AI is already embedded in everyday digital tools and not limited to standalone platforms.

To frame the policy as enabling safe innovation rather than appearing prohibitive, the intent should be explicit: to safeguard pupils and staff; ensure ethical and responsible use; support learning and workload efficiency; and to prepare pupils for an AI-enabled world.

Core principles

A strong policy is underpinned by clear principles with Safeguarding central to all decisions involving AI.

When using AI to make automated decisions that impact children and families, it is highly recommended that it is clearly communicated to the families which provides assurance that any decisions made are not reinforcing any bias, e.g. using AI to split children into classes.

Data protection is critical. Personal or sensitive information must not be entered into Gen-AI systems without the explicit consent of the Data Subjects and after conducting a Data Protection Impact Assessment (DPIA).

Applying critical thinking is key as outputs may be inaccurate, incomplete or biased and should always be verified. Human judgement remains essential to any outcome.

School approach to Gen-AI

The policy should define how Gen-AI will be used in context. This includes defining its purpose of use (i.e. workload reduction or learning support) and when Gen-AI can be used and when not.

It should also include who is permitted to use AI, any age or role-based restrictions, and expectations for responsible use.

It should also reflect reality as AI may already be in use informally. The goal is to bring its use into a safe, structured framework rather than eliminate it.

Stakeholder engagement is also important. Staff, pupils and parents may have differing views and levels of understanding. Involving them supports clarity and trust.

Alignment with existing policies

AI intersects with many existing policies and should not be stand alone. It should align with safeguarding and child protection; data protection legislation; behaviour and acceptable use policies, as well as assessment and examination policies. It should also align with staff conduct policies and complaints policies.

Data protection is particularly important. Schools must ensure transparency and compliance with data protection legislation, specifically around automated decision-making and profiling.

Roles and responsibilities

Clear accountability is essential. Governors and trustees provide oversight and ensure compliance. Senior leaders define strategy, approve tools and ensure consistent implementation. The Designated Safeguarding Lead manages safeguarding risks and incidents linked to AI use. Meanwhile, the Data Protection Officer ensures legal compliance and oversees DPIAs, and technical staff configure filtering, monitoring and system security.

However, all staff are responsible for using only approved AI tools, protecting personal sensitive data, checking AI outputs for accuracy and bias, modelling safe practice, and reporting concerns.

Without roles clearly defined, risks may increase and oversight may become challenging and fragmented.

Approval of Gen-AI tools

A formal approval process is essential for safe use. The policy should state that only approved Gen-AI tools may be used for school purposes and that new tools must not be introduced without permission and assessment.

To create a high benchmark, the DfE’s Gen-AI Product Safey Standards should be utilised when assessing any tools. Assessment considerations include data protection, in-tool filtering and monitoring, emotional attachment, cognitive offloading, and mental health.

It is recommended that the tools are reviewed regularly due to the ever-evolving changes in Gen-AI. Schools should consider how they make this list of tools available to all staff, including supply staff and visitors (e.g. a Social Worker wanting to use a Gen-AI note taking tool).

Filtering and monitoring

Filtering and monitoring are key safeguarding controls. Schools should confirm that gen-AI tools are blocked by default and approved tools are allowed by exception to certain user groups. They should also ensure that filtering is regularly tested and reviewed on all school owned devices (on and off site).

This is not a one-time setup. Gen-AI tools and risks can evolve quickly. Analysing monitoring data is a powerful tool, using them can help inform staff training, curriculum planning and safeguarding responses.

Reporting and recording

Clear reporting routes are essential. The policy should outline how to report safeguarding concerns involving AI and the misuse of AI tools, e.g. a pupil uses Gen-AI to complete a piece of work without permission or using Gen-AI to create harmful fake imagery. It should also include how to report data protection incidents. 

For issues to be addressed early, it should explain how incidents are recorded and escalated in-line with safeguarding and data protection procedures.

Teaching and learning

The policy should define expectations for artificial intelligence. For pupils, this means when use is permitted, how use should be acknowledged, expectations for independent work, and limits on over-reliance.

For staff, this should define appropriate use in planning and resource creation, boundaries around assessment and feedback, intellectual property considerations, and responsibility for checking outputs.

Developing AI literacy in pupils helps their knowledge and skills to enable them to critically understand, evaluate, and use AI systems and tools safely and ethically.

Staff support and review

Schools should provide ongoing support to use AI safely and effectively for staff, governors and trustees. This should include professional development on artificial intelligence, training on bias, data protection and ethical risks, as well as support in evaluating Gen-AI outputs critically. There should also be guidance for classroom application.

AI capability is not static, therefore it’s recommended that training is ongoing, rather than one-off.

LGfL AI policy toolkit

LGfL’s AI Policy Toolkit supports schools in developing a safeguarding-led approach to AI, explains what to include, key risks and how to align AI with existing expectations.
It is designed to be flexible rather than prescriptive, allowing schools to adapt it to their own context.

Please explore the toolkit and access self-service training here.

Conclusion

AI is already helping to shape education and by ignoring it, may increase the risk rather than eliminating it.

A clear AI policy helps to provide control, consistency and safeguarding assurance. It supports staff, protects pupils and ensures AI is used purposefully and responsibly. The goal is not to avoid AI, but to manage it well.