Keeping Children Safe in Education
Expects schools to respond to emerging online risks, including the misuse of images.
A photograph of a class on sports day tells a prospective parent more than any paragraph can. It tells a child that what they did mattered. That is why schools publish so many, and why the answer is not to publish fewer.
An identifiable photograph of a child, published online, can now be used to make images that never happened. The tools are free, they run on an ordinary computer, and they do not need a name or a caption. They need a face.
In July 2026 the Internet Watch Foundation and the National Crime Agency issued guidance after criminals took photographs of pupils from a UK school’s own website, generated more than a hundred sexual images of those children with AI, and tried to blackmail the school.
What the Internet Watch Foundation found in 2025
8,029
AI generated images and videos of child sexual abuse, up 14% on 2024
3,443
of them videos, against 13 the year before
Source: Internet Watch Foundation, Harm without limits, March 2026
Nothing was hacked. The photographs were published, as every school publishes them.
94%
say school photography matters to their school’s culture, identity and reputation
93%
are aware criminals are using AI to generate deepfake imagery from school websites and social media
49%
say their school has experienced pupil image misuse, manipulation or an extortion attempt
54%
are considering removing identifiable photographs of pupils altogether
Read the first and last together. More than half of heads are considering giving up something 94% say matters to who their school is. That is not a safeguarding decision. It is a school losing something because the only option it can see is less.
Censuswide survey of 500 UK headteachers, June 2026. Read the full findings in In Plain Sight.
No new policy is needed to take this seriously. The documents schools already work to have written the line.
Expects schools to respond to emerging online risks, including the misuse of images.
Requires every early years setting’s safeguarding policy to cover how cameras and image sharing devices are used.
Asks schools and parents to review image consent and who can see what is posted.
Asks activity providers to audit the photographs already on their websites.
Each of them asks a school to think about its photographs. None of them asks it to stop publishing them.
Consent decides whether a photograph can be published. It cannot reach a photograph once it is public. There is no known audience, no access control, and nothing to withdraw once an image has been copied.
Does the person looking at this photograph need to see the identity of the actual child?
For a website, a prospectus or a public social media post, the answer is almost always no.
Option one
Publish as before and accept the risk to every child whose face is online.
Option two
Remove identifiable photographs and lose the thing 94% of heads say matters.
Option three
Take the real child’s identity out of the picture before it goes online.
Aidos exists for the third option. It replaces each real face with a realistic face that belongs to no one, and leaves everything else in the photograph exactly as it was. It does not replace your photography policy, your consent process or your staff training. It answers one question, and leaves the rest where it is.
Start a no obligation 28 day free trial today with 15 free images. See how easily Aidos replaces real faces with realistic replacement faces, keeping pupils safe while your school photography stays natural and authentic.
Start trialWant to know more, or see how Aidos can help your school? Request a demo today.
Request a demo