Does the DfE's AI Strategy for Schools Go Far Enough to Address Safeguarding Concerns?

The DfE's AI strategy for schools does not go far enough to address safeguarding concerns; it leaves critical gaps that put pupils and staff at risk and fails to reassure parents or school leaders adequately.

Here's why I think this

In my experience working closely with UK schools through LESSO AI, I see a stark mismatch between the DfE's high-level AI guidance and the gritty reality of safeguarding challenges on the ground. The strategy leans heavily on existing frameworks like KCSIE 2025 but stops short of providing concrete, enforceable safeguards specifically tailored to AI's unique risks. Schools are expected to interpret broad principles into policy without clear, practical tools or accountability measures.

Parents, who are rightly anxious about AI's impact, remain largely unconvinced. Only 12% of parents support AI use in lessons, according to a recent Parentkind survey [3]. This distrust stems from a lack of transparent, robust safeguards that address data privacy, content moderation, and the potential for AI to amplify harmful biases or misinformation. The DfE's approach feels like a box-ticking exercise rather than a proactive shield.

Meanwhile, safeguarding teams in schools are overwhelmed. They face an influx of AI-generated content flagged by filters but lack guidance on how to assess risks or intervene effectively. The strategy's reliance on schools' existing safeguarding processes ignores the fact that AI introduces new, complex threats that require specialist understanding and resources.

What most people get wrong about AI safeguarding in schools

Many believe that existing safeguarding frameworks are sufficient to cover AI risks. The common assumption is that if schools comply with KCSIE 2025 and data protection laws, they are safe. This is misleading.

AI is not just another digital tool; it's a disruptive technology that can generate unpredictable content and interact with pupils in ways traditional software cannot. The nuances of AI-generated misinformation, deepfakes, or biased outputs demand a fundamentally different safeguarding mindset. Simply folding AI into existing policies without dedicated training, monitoring, and incident response strategies is naive.

Another misconception is that AI's benefits in education outweigh the risks, so safeguarding concerns can be managed later. This dangerously underestimates the immediacy and scale of potential harm, from exposure to inappropriate content to manipulation through AI chatbots. The DfE's strategy tacitly endorses this complacency by prioritising innovation over protection.

What I've seen actually happen in schools

From direct conversations with headteachers and safeguarding leads, I've witnessed a growing anxiety about AI's unchecked use. Many schools have hastily drafted AI policies that mimic the DfE's vague guidance but fail to address real-world scenarios like:

• Pupils using AI to generate harmful or plagiarised content undetected by staff
• Insufficient filtering of AI chatbots that can produce inappropriate or biased responses
• Staff uncertainty on how to incorporate AI risk flags into safeguarding workflows
• Parental backlash due to lack of communication and transparency about AI use

Despite 89% of schools using the GOV.UK website for mandatory data returns [4], fewer than half have a comprehensive AI safeguarding policy in place. This disconnect means many schools are flying blind, relying on reactive rather than preventative measures.

The DfE's AI strategy feels like a box-ticking exercise rather than a proactive shield against real safeguarding risks.

What I'd do differently

First, I would demand that the DfE develop a dedicated AI safeguarding framework that goes beyond existing policies. This framework must include:

• Clear, enforceable standards for AI content moderation tailored to school contexts
• Mandatory training for safeguarding leads on AI-specific risks and incident management
• Transparent communication protocols to engage parents and pupils about AI use and safety
• A centralised reporting and support system for AI-related safeguarding incidents

Second, the strategy should require AI vendors to meet strict safety certifications before schools can adopt their products. This would shift some responsibility away from overstretched schools and ensure safer AI tools by design.

Finally, I would push for ongoing research and data collection on AI's impact in schools to inform policy updates. The current strategy's static guidance risks becoming obsolete as AI technology rapidly evolves.

The uncomfortable truth

The uncomfortable truth is that the DfE's cautious, minimalist approach to AI safeguarding is less about protecting children and more about avoiding regulatory burdens on schools and vendors. The government's desire to encourage innovation has overshadowed the urgent need for robust safety nets.

This means schools are left to fend for themselves, often without the expertise, resources, or clarity they need. The consequence? Increased risk of harm to pupils and reputational damage to schools. Some school leaders might even choose to ban AI outright, missing out on legitimate educational benefits because the safeguarding framework is inadequate.

If we don't confront this reality head-on, the gap between policy and practice will widen, and safeguarding failures will become inevitable.

I'm keen to hear from school leaders and safeguarding professionals about your experiences with AI in your settings. Do you feel the DfE's strategy supports you, or are you left to navigate these risks alone? Let's start a conversation — share your thoughts and stories on LinkedIn or reach out directly. The future of safe AI in schools depends on honest dialogue and collective action.

Generate Your School's AI Policy — Free

Mrs J drafts a UK-compliant AI policy for your school in under a minute. Mapped to DfE guidance, DUAA 2025, and KCSIE 2025. Sent directly to your inbox.