AI, School Websites and Images!

[andy_b]

NEW Guidance about publishing photographs of children online - (UKSIC)  Schools rightly want to celebrate the achievements of their pupils, and photographs are a natural part of that. But new guidance from the UK Safer Internet Centre sets out why that instinct now needs to be balanced carefully against an emerging safeguarding risk. The guidance follows a confirmed incident, reported to the Internet Watch Foundation by UK police, in which criminals harvested photographs from a secondary school's website, used AI tools to turn them into child sexual abuse material, and demanded money from the school with the threat of wider distribution. Incidents of this kind are not yet widespread, but the working group behind the guidance is clear that it is only a matter of time before more schools are targeted. The practical steps are straightforward: avoid face-on, identifiable photographs, never pair names with images, remove EXIF metadata before uploading, and publish only low-resolution versions. Existing content on your website and social media is worth reviewing too. The question is not whether to celebrate pupils, but how to do so without creating unnecessary risk. The full guidance, including a checklist of actions, is available here: https://saferinternet.org.uk/guide-and-resource/image-guidance-for-education-settings

[DavR]

2 hours ago, andy_b said:

The guidance follows a confirmed incident, reported to the Internet Watch Foundation by UK police, in which criminals harvested photographs from a secondary school's website, used AI tools to turn them into child sexual abuse material, and demanded money from the school with the threat of wider distribution.

 

Jeez, just when you thought you'd seen it all, technology enables a new low in human behaviour.

 

What a terrible day to have eyes.

[synaesthesia]

Yup, said this to our DSL earlier. 

Schools unable to use images of our entire clientbase (ugh, sorry!) in any promotion because some ***** keep pulling stunts like this. Sick people.

Unfortunately inevitable I suppose with AI as it is

[3s-gtech]

Hmmmm. Do we generate generic images for school media using AI generated students or at least their faces? Also deals with issues where a parent/student withdraws consent.

[synaesthesia]

Where do those AI generated images come from?

Amalgamations of images they've already chewed - for all I know, my face has already been consumed and spat out in some amalgamation of big footed balding fat bloke

Prospectus' could look rather crap going forward.

Of course this isn't a ban - it's be more careful, monitor more often but even then I wonder what's the point - once that image has gone onto the internet, it's there forever even after you've expunged it from your own website. 

[Rob_D]

Pixel art website images? Actually, I think that looks kind'a cool.😁

[3s-gtech]

"Generate images of our school using the cast of 1970s Grange Hill, but gender-swap them all and make their eyes slightly too far apart"

 

Sorted. Disclaimer: you may lose prospective students.

[andy_b]

25 minutes ago, 3s-gtech said:

Hmmmm. Do we generate generic images for school media using AI generated students or at least their faces? Also deals with issues where a parent/student withdraws consent.

[mullet_man]

3 hours ago, Rob_D said:

Pixel art website images? Actually, I think that looks kind'a cool.😁

It does say to use low resolution images! 

[BKGarry]

4 hours ago, Rob_D said:

Pixel art website images? Actually, I think that looks kind'a cool.😁

Can I see Guybrush Threepwood in that image!!!

[synaesthesia]

Still have to be careful, could they be individually identifiable?

I mean, many of us would be able to pick out at least half of these:

 

[Jobos]

The HT looks after the website so had a quick look and following the guidelines would mean 3/4 of the photos would be removed. This is the same old advice just reworded to take into account AI and I always think how many parents who reject having photos of their children on on the school website would jump at the chance of having the same photo in the local rag.

 

[mcrompton01]

Hi all,

 

Long time lurker, not quite first time posting. I safeguarding lead brought this to our attention and I'm trying to find out if other schools or Trusts have issued any guidance or advice to its staff or modified their websites etc to reduce the risk. 

 

UK schools should remove pupils’ online photos as AI blackmail threat grows, say experts | Internet safety | The Guardian

 

We are having a meeting on Monday around a response for example, take images side on, group shots, at a distance if possible, use image on websites more focused on an activity not the person. 

 

Thanks in advance.  

[lparry]

I know I shouldn't be surprised but how is this the timeline we now live in

 

[TechMonkey]

One of the consequences of this could be more parents withdrawing photo consent, which if it hits a threshold, makes it very hard to use images at all.  Suddenly any marketing or social media strategy collapses (oh no!) and a new one has to be created.

This did come up about a year ago and I know it was discussed here, with the suggestions being along the lines of what has been suggested.  There are systems that genericify people, but is it enough for a school to say that porn created from website photos isn't a real person, or is the damage done by having the school's uniform used in these pictures?

 

It's a tricky situation as how do you account for scummy things other people will do, and how far do you go to prevent or counter it.  As LParry says, if people could stop being scummy it would solve a lot of problems.

[Rob_D]

This was also being discussed here: 

 

[Dos_Box]

I'll merge the threads.

[ConceroEdu_Matt]

What a time to be alive. 

 

I guess, if people are agreeing to put their or their childrens pictures on the web, then you sort of have to accept these risks these days. Key thing is to communicate them, and do what mitigation you can.

[Rob_D]

Has anyone looked into using one of these anti-ai things like Glaze? https://glaze.cs.uchicago.edu/what-is-glaze.html

 

[Simcfc73]

I've been asked about this earlier this morning. Looks like we will be replacing all our images,

 

Tried this but didn't recognise the faces on group pics.

 

Safe Face AI

Edited May 22 by Simcfc73

[Rob_D]

4 minutes ago, Simcfc73 said:

I've been asked about this earlier this morning. Looks like we will be replacing all our images,

 

What sort of alternatives are you going with?

[Max_Power]

We have considered Aidos Protects – Protecting schools and students from deepfake exploitation. 

 

While it offers some mitigation, on the flip side, doing this type of thing can remove the personal / individual recognition of pupils achievements and celebrations we post on X fairly regularly.

 

 

[djm968]

On 12/05/2026 at 08:58, Jobos said:

The HT looks after the website so had a quick look and following the guidelines would mean 3/4 of the photos would be removed. This is the same old advice just reworded to take into account AI and I always think how many parents who reject having photos of their children on on the school website would jump at the chance of having the same photo in the local rag.

 

Actually, if you read the guidance properly, it isn’t just “the same old advice reworded to take into account AI.” It’s been shaped by a real case where a school was threatened with the release of inappropriate AI‑generated images created using photos of pupils taken from the school website and social media.

As Jobos pointed out, applying the guidance strictly could mean removing a large proportion of website photos. But the key difference here is the risk landscape has changed — this isn’t just about general privacy concerns anymore, it’s about how easily images can now be scraped and misused with AI tools.

 

I do understand the frustration, and the comparison with parents being more relaxed about photos in local newspapers is a fair one. However, digital images on publicly accessible websites are far more exposed, easier to collect at scale, and far more vulnerable to manipulation than a one-off print publication.

So while some of the advice may feel familiar, the context and potential consequences are very different now.

 

https://saferinternet.org.uk/blog/schools-given-advice-on-image-safety-to-keep-ahead-of-threat-from-ai-blackmailers

Edited May 22 by djm968
Added link

[Olliedawg]

Our SLT are discussing this at the mo.. 

 

I had a thought about running the student photos on our website through CoPilot within our 365 tenant (ironic, use AI to beat AI..) to "cartoonify" the images, making the faces non recognisable.

[steveg]

We had a meeting about this this week.  Our website and social media will have faces changed by ai now so real looking but not actually real children.  We are going to start emailing out newsletters again (rather than doing them through our website) and have requested a new consent from parents to allow images of their children to be shared with other parents only via email. 

 

This seemed the only logical conclusion.  We asked for consent for images for social media and website previously, and in light of the new guidance would need to re-request this.  We were going to share the new information, and then re-request permission, but then realised the email would go something like: We've had new guidance that there is now a real and proven risk of us having your children's images on our website where bad things may happen to them and we may be blackmailed by them.  Do you give us permission to put them on there?  We couldn't see a lot of parents saying yes to this, so concluded we should just remove them.