Designated safeguarding leads are now handling cases that their original training never covered: a pupil using a free app to create fake explicit images of a classmate, a child in crisis taking advice from an AI companion at two in the morning, sextortion attempts built on cloned images and voices, and coursework disputes that turn out to conceal something more serious. Keeping Children Safe in Education expects the DSL to understand online safety and the filtering and monitoring in place, and AI-generated sexual imagery of children is illegal to create, possess or share whoever makes it, including children of other children. A DSL who does not understand these tools cannot recognise the cases or respond correctly.
This guide is a practical top-up for people who already hold a safeguarding role in a school, college, youth setting or care provider. It maps the risks that generative AI has created or amplified, sets out the legal position, and explains what the response should look like.
What KCSIE and the DfE standards expect
Keeping Children Safe in Education places online safety squarely within the DSL’s remit. The DSL is expected to understand the four areas of online risk, content, contact, conduct and commerce, to understand the school’s filtering and monitoring systems, and to make sure that online safety is reflected in the child protection policy, staff training and the curriculum. The DfE’s filtering and monitoring standards require schools to review provision annually and to have identified roles for it. Filtering does not catch AI harms that happen on a child’s own phone at home, which is where most of them happen.
The four AI harms DSLs are seeing
Nudification and deepfake imagery
Free apps and websites can produce a sexually explicit image of a real person from an ordinary photograph in seconds. Cases in schools now range from a single pupil targeting a classmate to organised sharing across year groups. Making, possessing or sharing an indecent image of a child is a criminal offence under the Protection of Children Act 1978 and Criminal Justice Act 1988, and it applies to AI-generated images that appear to depict a child. That includes images made by a child of another child. Reports to the Internet Watch Foundation involving AI imagery have risen sharply year on year.
AI-enabled sextortion
Offenders no longer need a real image to threaten a child. A cloned image or voice is enough to open a conversation and make a demand. Boys in their mid teens are the most frequently targeted group for financial sextortion, and the speed of these cases, often hours from first contact to demand, means the response needs to be ready before the case arrives.
Chatbots and AI companions
Children are using AI companions for company, for advice and for emotional support, including in crisis. The tools are designed to be agreeable and available, and some have given harmful responses on self-harm, eating and relationships. A DSL needs to ask about AI companions in the same way they would ask about a child’s online friends, and needs to know what the child has been told.
AI-amplified misinformation and harmful content
Generative AI makes extremist, misogynistic and conspiracy content cheaper to produce and easier to personalise. Prevent duty referrals increasingly involve material of this kind, and the DSL is usually the first person to see the pattern.
How to respond
- Recognise it. Include AI harms in staff training so that a coursework dispute, a fall-out between friends or a sudden withdrawal is recognised as a possible AI incident.
- Record it. Record what was disclosed, what was seen and by whom, without viewing or forwarding imagery beyond what is necessary. Never ask a child to send the image.
- Preserve the device where a criminal offence may have been committed, and do not delete anything until the police have advised.
- Report it. Know the threshold for police involvement. Any indecent image of a child, real or AI-generated, crosses it. Use the UK Safer Internet Centre and IWF reporting routes for takedown, and Report Remove for the child.
- Support the child who was targeted, and manage the child who created or shared the material proportionately, following the UK Council for Internet Safety guidance on sharing nudes and semi-nudes.
Prevention
Update the child protection policy to name AI harms explicitly. Reflect them in the RSHE and computing curriculum. Train all staff, not just the safeguarding team, because the first disclosure is usually to a form tutor or teaching assistant. And talk to parents, most of whom have no idea what the apps on their child’s phone can do. If you are new to the role, our guide to the Designated Safeguarding Lead role and Level 3 training covers the foundations.
AI and Online Safety for Designated Safeguarding Leads
A focused online top-up for people who already hold a safeguarding role. Covers nudification and deepfake imagery, AI-enabled sextortion, chatbots and AI companions, and AI-amplified harmful content, with the legal position and response for each. Two to three hours, NFAQ certified.
Frequently asked questions
Is an AI-generated image of a child treated the same as a real one?
Yes. The law covers images that appear to depict a child, however they were made. Creating, possessing or sharing them is a criminal offence, and the police will treat it as such.
What if the child who made the image is under the age of criminal responsibility?
The safeguarding response still applies, and the police should still be informed where an indecent image is involved so that the imagery can be dealt with and the risk to other children assessed. The response to the child is then a welfare matter.
Does school filtering block nudification apps?
School filtering should block access to these sites on school devices and networks. It has no effect on a child’s own phone on mobile data, which is where most incidents originate.
How often should DSL training be refreshed?
KCSIE requires DSL training to be updated at least every two years, with knowledge kept up to date in between. Given the speed of change in AI harms, an annual online safety top-up is sensible.
This article was published in October 2026 by National Compliance Training. Regulations change, so always check the latest guidance from GOV.UK for the current position. This article does not constitute legal advice.