A child does not need to have taken a sexual image for an image to be used against them.
Artificial intelligence can place a child’s face into something that never happened. It can alter an ordinary school photograph, generate a convincing likeness or create a sexualised image from almost nothing. It can then be used to humiliate, threaten, blackmail or control.
The image may be fake. The fear of walking into school after it has been shared is not. The messages are not. The laughter, threats and feeling that everybody is looking are not.
This is where the quality of the first professional response becomes critical.
Here is a real truth: The first professional response can either return some control to the child or take more of it away.
The technology is new. The use of shame, sexual humiliation, coercion and peer status is not.
AI image abuse can sit inside child-on-child abuse, bullying, sexual harassment, exploitation, extortion, domestic abuse, adult offending or a wider pattern of control. Sometimes the image is circulating. Sometimes the child has only been threatened with it. Sometimes there is a rumour that an image exists and the social harm has already begun. That complexity is why a simple instruction to confiscate a phone or report an account is not enough. The safeguarding response must address the child’s whole experience: what has happened, who may be involved, where the material may be travelling, whether threats are active and what the next hour or next school day now feels like.
A calm first response should reduce shame, establish safety and keep the child involved.
The reassurance that can accidentally minimiseAn adult may instinctively say, “But it is not really you.” The intention may be to reassure. The child may hear something very different: it is not real, so it does not count.
A better response separates the truth of the image from the truth of the impact:
Try saying: “Even if the image was created or altered by AI, what has happened to you matters. I am not going to treat this as less serious because the image is not real.”
That sentence does three things. It believes the impact. It avoids pretending the adult already knows every fact. It tells the child that the response will not depend on whether the technology produced a perfect fake.
First response is not investigation
When professionals are anxious, they often try to close every gap immediately.
Who made it? Exactly what can be seen? How many people received it? Can you show me the chat?
Those questions can turn a disclosure into an interview and expose the child to repeated viewing. They can also lead staff to create or retain copies they should never have held.
The first response has a narrower job. It should establish immediate safety, identify active threats or extortion, place responsibility with the safeguarding lead and gather only enough information to decide the next action.
Safer first-response questions versus questions to avoid:
What are you most worried might happen today?
not
Why did you let this happen?
Is anyone threatening you or asking for money, images or contact?
not
Show me every chat and send me the image.
Where do you believe it has been shared?
not
Exactly what can you see in it?
What are you worried adults might do that could make this worse?
not
Do you want us to do nothing?
Preserve information, not extra copies
The safest evidence response is disciplined. Record the child’s words, platform or group-chat name, usernames, dates, times, URLs, threats and decisions where this can be done without opening, copying or forwarding the image.
Do not ask a child to send or download it. Do not print it, save it to a case file or forward it to prove the concern. Do not conduct an informal search of the child’s device. Where police involvement is likely, seek advice about preservation and device handling before altering anything.
Removal remains important, but timing and method need coordination.
Report Remove can help an under-18 in the UK report nude or sexual imagery of themselves, including imagery created or altered using AI. The setting should not keep its own copy to support the request.
Safety cannot look like punishment
A child-centred plan considers the next hour, the next journey, the next lesson and the next online interaction. It asks who will meet the child, how they will move through the setting, what peer contact needs managing, what learning adjustments are wanted and when the plan will be reviewed.
It also asks a harder question: have we restricted the targeted child because that is administratively easier?
Removing the child from lessons, friendship groups, activities or their device may reduce immediate pressure. If every restriction lands on the child who was targeted, the safeguarding response can begin to feel like another punishment.
The alleged creator or sharer may need two responses
A child alleged to have created or shared a sexualised image may have caused serious harm. Clear boundaries and accountability are required.
That same child may also be acting under coercion, copying peer behaviour, experiencing exploitation or lacking the developmental understanding adults assume. Safeguarding language should not disappear simply because a child is accused of causing harm.
Holding those two truths is not being soft. It is how settings protect every child without collapsing harm, intent, risk and need into a single judgement.
Policy will not speak for staff under pressure
Keeping Children Safe in Education 2026 explicitly recognises intimate images and videos generated using AI within child-on-child abuse. UKCIS guidance also covers AI-generated imagery and sets a clear expectation that staff should not view imagery, with exceptional viewing decisions reserved for the DSL or equivalent where necessary to safeguard the child. But policy is only the beginning.
A document cannot decide what a teaching assistant says when a frightened child tries to hand over a phone. It cannot rehearse the night worker’s response when a child in residential care discloses after hours. It cannot ensure that parent contact is calm, or that tomorrow’s return plan does not isolate the child further.
Staff need usable language, clear boundaries and a rehearsed route. They need to know what to notice, what to say and what not to do.
The question for every setting: How confident are the adults in your setting that they would know what to say first?
Start with the first hour
Smart Training & Consultancy has developed an AI Image Abuse First-Response range for schools, colleges and children’s homes:
Download the First-Hour Guide here for free,
Purchase the full practitioner kit here
Or go purchase the full staff implentation pack here
or alternatively book a staff briefing at smarttc.co.uk.
For organisational enquiries, email [email protected].
Sources and further reading
Artificial intelligence can place a child’s face into something that never happened. It can alter an ordinary school photograph, generate a convincing likeness or create a sexualised image from almost nothing. It can then be used to humiliate, threaten, blackmail or control.
The image may be fake. The fear of walking into school after it has been shared is not. The messages are not. The laughter, threats and feeling that everybody is looking are not.
This is where the quality of the first professional response becomes critical.
Here is a real truth: The first professional response can either return some control to the child or take more of it away.
The technology is new. The use of shame, sexual humiliation, coercion and peer status is not.
AI image abuse can sit inside child-on-child abuse, bullying, sexual harassment, exploitation, extortion, domestic abuse, adult offending or a wider pattern of control. Sometimes the image is circulating. Sometimes the child has only been threatened with it. Sometimes there is a rumour that an image exists and the social harm has already begun. That complexity is why a simple instruction to confiscate a phone or report an account is not enough. The safeguarding response must address the child’s whole experience: what has happened, who may be involved, where the material may be travelling, whether threats are active and what the next hour or next school day now feels like.
A calm first response should reduce shame, establish safety and keep the child involved.
The reassurance that can accidentally minimiseAn adult may instinctively say, “But it is not really you.” The intention may be to reassure. The child may hear something very different: it is not real, so it does not count.
A better response separates the truth of the image from the truth of the impact:
Try saying: “Even if the image was created or altered by AI, what has happened to you matters. I am not going to treat this as less serious because the image is not real.”
That sentence does three things. It believes the impact. It avoids pretending the adult already knows every fact. It tells the child that the response will not depend on whether the technology produced a perfect fake.
First response is not investigation
When professionals are anxious, they often try to close every gap immediately.
Who made it? Exactly what can be seen? How many people received it? Can you show me the chat?
Those questions can turn a disclosure into an interview and expose the child to repeated viewing. They can also lead staff to create or retain copies they should never have held.
The first response has a narrower job. It should establish immediate safety, identify active threats or extortion, place responsibility with the safeguarding lead and gather only enough information to decide the next action.
Safer first-response questions versus questions to avoid:
What are you most worried might happen today?
not
Why did you let this happen?
Is anyone threatening you or asking for money, images or contact?
not
Show me every chat and send me the image.
Where do you believe it has been shared?
not
Exactly what can you see in it?
What are you worried adults might do that could make this worse?
not
Do you want us to do nothing?
Preserve information, not extra copies
The safest evidence response is disciplined. Record the child’s words, platform or group-chat name, usernames, dates, times, URLs, threats and decisions where this can be done without opening, copying or forwarding the image.
Do not ask a child to send or download it. Do not print it, save it to a case file or forward it to prove the concern. Do not conduct an informal search of the child’s device. Where police involvement is likely, seek advice about preservation and device handling before altering anything.
Removal remains important, but timing and method need coordination.
Report Remove can help an under-18 in the UK report nude or sexual imagery of themselves, including imagery created or altered using AI. The setting should not keep its own copy to support the request.
Safety cannot look like punishment
A child-centred plan considers the next hour, the next journey, the next lesson and the next online interaction. It asks who will meet the child, how they will move through the setting, what peer contact needs managing, what learning adjustments are wanted and when the plan will be reviewed.
It also asks a harder question: have we restricted the targeted child because that is administratively easier?
Removing the child from lessons, friendship groups, activities or their device may reduce immediate pressure. If every restriction lands on the child who was targeted, the safeguarding response can begin to feel like another punishment.
The alleged creator or sharer may need two responses
A child alleged to have created or shared a sexualised image may have caused serious harm. Clear boundaries and accountability are required.
That same child may also be acting under coercion, copying peer behaviour, experiencing exploitation or lacking the developmental understanding adults assume. Safeguarding language should not disappear simply because a child is accused of causing harm.
Holding those two truths is not being soft. It is how settings protect every child without collapsing harm, intent, risk and need into a single judgement.
Policy will not speak for staff under pressure
Keeping Children Safe in Education 2026 explicitly recognises intimate images and videos generated using AI within child-on-child abuse. UKCIS guidance also covers AI-generated imagery and sets a clear expectation that staff should not view imagery, with exceptional viewing decisions reserved for the DSL or equivalent where necessary to safeguard the child. But policy is only the beginning.
A document cannot decide what a teaching assistant says when a frightened child tries to hand over a phone. It cannot rehearse the night worker’s response when a child in residential care discloses after hours. It cannot ensure that parent contact is calm, or that tomorrow’s return plan does not isolate the child further.
Staff need usable language, clear boundaries and a rehearsed route. They need to know what to notice, what to say and what not to do.
The question for every setting: How confident are the adults in your setting that they would know what to say first?
Start with the first hour
Smart Training & Consultancy has developed an AI Image Abuse First-Response range for schools, colleges and children’s homes:
- A free two-page First-Hour Guide for immediate staff use
- A full practitioner kit with child and parent scripts, evidence guidance, safety planning, peer response, supervision tools and templates
- A 60-minute staff briefing using realistic scenarios and local procedure mapping
- Organisational support for policy-to-practice review and wider safeguarding development
Download the First-Hour Guide here for free,
Purchase the full practitioner kit here
Or go purchase the full staff implentation pack here
or alternatively book a staff briefing at smarttc.co.uk.
For organisational enquiries, email [email protected].
Sources and further reading
- Keeping Children Safe in Education 2026
- UKCIS: Sharing nudes and semi-nudes advice for education settings
- NCA: guidance on AI-manipulated images of children, July 2026
- IWF: AI child sexual abuse imagery parent safety guide
- Childline and IWF: Report Remove
- Crime and Policing Act 2026: child sexual abuse material factsheet
RSS Feed