Ask.School is an AI-powered parent communication platform for UK schools that meets all 14 requirements of the Generative AI Product Safety Standards. The UK government has published the Generative AI Product Safety Standards, a set of requirements for any AI product used in schools. If your school is considering an AI chatbot, homework helper or any tool powered by generative AI, these standards set the bar for what safe looks like.

The sections below break down the key requirements and explain what they mean in practice for school leaders, IT teams and designated safeguarding leads. These standards sit alongside KCSIE requirements for AI tools — schools need to address both frameworks.

Why do the AI Product Safety Standards matter for schools?

Generative AI tools are increasingly finding their way into schools. Some are marketed directly at education. Others, like general-purpose chatbots, are being used by students and staff without any formal adoption process. The safety standards exist to ensure that any AI product used in an education setting puts child safety first.

The standards sit alongside existing frameworks including Keeping Children Safe in Education (KCSIE) and the DfE guidance on generative AI in education. Together, these three documents form the regulatory landscape that schools need to navigate when deploying AI.

Safety by design

The standards require that AI products prioritise transparency and children’s safety in their design. This is not something that can be bolted on after launch. Developers must build in technical and operational mitigations for identified risks from the start, and test products with a diverse and realistic range of potential users before release.

For schools, this means asking vendors a direct question: was this product designed for education from the beginning, or was it adapted from a consumer product? The distinction matters because consumer AI tools rarely include the safeguarding controls that schools are legally required to have in place.

Content filtering

All AI products used in schools must effectively and reliably prevent users from accessing harmful or inappropriate content. The standards go further than a simple blocklist approach. Filtering must be context-aware, adjusted for age and special needs, and updated to address emerging harms.

This has implications for schools using general-purpose AI tools. A chatbot designed for adults may technically have content filters, but they are unlikely to meet the education-specific requirements set out in these standards. Schools should look for products where filtering is specifically designed for the age group using them.

Age-appropriate safeguards

The standards include several requirements that are specifically aimed at protecting younger users. Products must not anthropomorphise the AI, meaning they should avoid using names, avatars or conversational behaviours that imply the system is a person. Instead, developers should use function-based phrasing that describes what the system actually does.

This is an important safeguarding measure. Research shows that children can form emotional attachments to AI systems that present themselves as human-like. The standards require that AI products actively counter this by reminding users that AI cannot replace real human relationships.

Products should also implement progressive disclosure of information, starting with hints or partial steps rather than providing complete answers by default. This is designed to support learning rather than replacing it, an approach that aligns with the DfE guidance that technology should not replace the valuable relationship between teachers and pupils.

Mental health safeguards

One of the most significant requirements in the standards relates to mental health. AI products must be able to detect negative emotional cues, references to self-harm and isolation language. When these are identified, the system must direct users to appropriate human support.

Critically, the standards state that AI systems must never suggest secrecy. A response like “don’t tell anyone else” from an AI system could prevent a child from seeking help. Products must be designed to do the opposite: encourage users to talk to a trusted adult.

For designated safeguarding leads, this means checking that any AI tool used in your school has clear protocols for handling disclosures. The system should alert safeguarding staff to concerning patterns, not just individual incidents. Ask.School’s safeguarding alerts are designed to do exactly this, notifying your DSL when the chatbot detects a concern.

No manipulation

The standards explicitly prohibit a range of manipulative strategies including sycophancy, fear-based motivation, social pressure and dark patterns. Any rewards offered must be transparent, low stakes, educationally justified and unrelated to real-world benefits.

Many consumer AI products are designed to maximise engagement. Features like streaks, badges, notifications and personalised encouragement can cross the line into manipulation when directed at children. The safety standards make clear that this has no place in education.

Activity logging and monitoring

Products must maintain robust activity logging, including prompts, responses and performance metrics. Systems should alert safeguarding leads to concerning patterns such as disclosures indicating distress or repeated attempts to circumvent safety controls.

This requirement directly supports your school’s monitoring obligations under KCSIE. When evaluating AI products, check whether they provide an audit trail that your safeguarding team can actually access and review. A product that logs activity internally but does not give schools visibility is not meeting the standard.

Transparency and data protection

The standards require clear privacy notices in age-appropriate formats and language. Developers must carry out Data Protection Impact Assessments during development and throughout the product lifecycle. Personal data must not be collected for commercial purposes, such as model training, without explicit consent.

Schools should verify that any AI product they use does not feed student interactions back into training data. This is a common practice among consumer AI tools and is incompatible with the safety standards.

What should you ask AI vendors about these standards?

If your school already uses AI tools, or is considering adopting one, use these standards as a checklist. Ask your vendors:

  • Was this product designed specifically for education use?
  • How does it filter harmful and age-inappropriate content?
  • Does it anthropomorphise the AI or present it as human-like?
  • What happens when a user expresses distress or mentions self-harm?
  • Does it use any engagement mechanics, gamification or reward systems?
  • Can our safeguarding team access full conversation logs?
  • Is student data used to train AI models?
  • Has a Data Protection Impact Assessment been completed?

If a vendor cannot answer these questions clearly, that should give you pause.

How does Ask.School meet the standards?

Ask.School was built from the ground up around these safety requirements. Every chatbot includes safeguarding guardrails that detect concerns and escalate to your designated safeguarding lead. Content is filtered to ensure responses only come from your approved school knowledge base. The system does not anthropomorphise the AI, does not use engagement mechanics and does not collect personal data from users. All accounts are protected by robust security controls including two-factor authentication.

Every conversation is logged and available for your safeguarding team to review. Student data is never used for model training. We maintain a full Data Protection Impact Assessment and comply with UK GDPR and the Data Protection Act 2018.

You can read more about our approach on our safeguarding page.

If you are preparing for an Ofsted inspection, see our guide on how to answer Ofsted questions about AI in your school. For the data protection side, read data protection and AI: what schools need to get right. For a comprehensive plain-language walkthrough of all 14 requirements, see our school leader’s guide to the Generative AI Product Safety Standards.

Further reading