DSLs, governors and staff doing safeguarding training keep asking the same question: when did AI first appear in statutory safeguarding guidance, and which documents now matter? Here is the timeline in plain English.
The short answer
Keeping Children Safe in Education (KCSIE) 2025, in force from 1 September 2025, was the first edition of the statutory safeguarding guidance to explicitly reference generative AI. It did so in its online safety section on filtering and monitoring, by adding a link to the DfE's Generative AI: product safety expectations. It was a signpost rather than a new standalone duty — but it brought AI into statutory safeguarding guidance for the first time.
KCSIE 2026, in force from 1 September 2026, builds on that and deals with AI more directly. See our summary of the KCSIE 2026 AI changes.
The timeline
| When | Document | What it did for AI |
|---|---|---|
| 2023 | DfE policy paper: Generative AI in education | Set out the DfE's position on generative AI in schools, including data protection and the need to protect pupils. Updated since. |
| 2023 | DfE filtering and monitoring standards; KCSIE 2023 | Made the DSL lead for filtering and monitoring and set standards schools must meet — the framework AI later slots into. |
| January 2025 | DfE: Generative AI: product safety expectations | Set out what AI products used in education should do to be safe. |
| September 2025 | KCSIE 2025 | First explicit AI reference in statutory safeguarding guidance, linking to the product safety expectations under filtering and monitoring. |
| 2026 | Updated product safety expectations | Expanded the harms AI products should guard against, including effects on learning, wellbeing and manipulation. |
| September 2026 | KCSIE 2026 | Deals with AI more directly — see our KCSIE 2026 summary. |
Guidance changes, so always check the current versions on GOV.UK.
Want a straight answer for your school?
Ask me anything about AI in your school — policy, tools, training or safeguarding. I reply personally, usually the same day.
Why this matters for schools
The practical consequence is that AI is now part of safeguarding, not just IT. In particular:
- Filtering and monitoring reviews should consider AI tools and how pupils access them. See KCSIE filtering and monitoring for AI.
- AI tools used with pupils should be checked against the product safety expectations.
- Online safety content risks — including misinformation and AI-generated imagery — belong in staff training. See deepfakes: what DSLs need to know.
- Your safeguarding policy should reference AI. See updating your KCSIE policy for AI.
Quick quiz answers
If you are revising for safeguarding training, the key points are:
- First statutory safeguarding reference to AI: KCSIE 2025, via the link to the product safety expectations.
- Who leads on filtering and monitoring: the DSL, working with IT.
- How often filtering and monitoring should be reviewed: at least annually.
- The move from signposting to more direct expectations: KCSIE 2026.
Other documents worth knowing
- DfE: Using AI in education support materials for school and college leaders and staff.
- JCQ guidance on AI in assessments — for exam and coursework integrity. See our JCQ explainer.
- ICO guidance on AI and data protection, and the Children's code.
- Plan technology for your school — the DfE service schools can use to check themselves against technology standards, also linked from KCSIE.
What to update this term
- Check your safeguarding policy references AI and the current KCSIE.
- Make sure your filtering and monitoring review includes AI tools.
- Include AI risks — deepfakes, chatbots, misinformation — in staff safeguarding training.
- Brief governors on what has changed.
Frequently asked questions
Which document first explicitly referenced AI within statutory safeguarding guidance?
Keeping Children Safe in Education 2025, in force from 1 September 2025, which linked to the DfE's Generative AI: product safety expectations in its filtering and monitoring section.
Does KCSIE 2026 mention AI?
Yes. KCSIE 2026, in force from 1 September 2026, builds on the 2025 reference and deals with AI more directly.
What are the DfE generative AI product safety expectations?
A DfE document, first published in January 2025 and since updated, setting out what AI products used in education should do to be safe, including filtering, monitoring, security and data protection.
Who is responsible for AI in school safeguarding?
The DSL takes lead responsibility for filtering and monitoring, working with IT staff, with governors providing oversight.
Where can I find the current guidance?
On GOV.UK. Guidance is updated regularly, so always check the current version.