Asking an AI tool to restyle a photo you took yourself is not the same as stripping a photographer's watermark, which copies their work without paying. The first costs nothing but sends your child's picture to a company that may keep it and, on a standard ChatGPT account, train on it unless you switch that off. If the school needs an official photo, its rules come first.
Last updated October 19, 2026.
What are parents actually doing with AI and school photos?
Two different things, and the coverage tends to blur them.
The first is a home studio. A parent photographs their children against a plain wall, uploads the pictures to ChatGPT and asks for a school-portrait look, usually a generic backdrop. Futurism described the trend on October 6, 2026, and Gadget Review reported the same day that school photo packages commonly cost $25 to $60 per child, which is what the trick is meant to avoid.
The second is watermark removal. The Week reported on October 8, 2026 that some parents use AI to remove the photographer's watermark from the preview images sent home, then print the result without paying.
Both are in the news in the US and the UK this month. I am treating them separately because the answer to "is this okay?" is different for each.
Is it illegal to use AI on a school photo?
Removing a watermark and restyling your own photo are different acts, so the law touches them differently. This is my reading of public guidance, not legal advice.
Photographs are protected by copyright automatically. The UK government's guidance says that copyright stops others from copying your work, making an adaptation of it or putting it on the internet without permission, and lists photography among the works covered. The US Copyright Office says copyright "exists from the moment the work is created" and names photographs among the protected works. A preview image sent home by a school photographer is their work. Running it through an AI tool to remove the watermark is copying and adapting it, and that is what the photographers are objecting to. Photographer Victoria Welton told Sky News, as PetaPixel reported on October 14, 2026, that the practice is illegal. She is a working photographer with an interest in the answer, and the PetaPixel page names no specific law.
A photo you took yourself is a different case. You made it, so editing it is not copying someone else's work. The question there is not the law. It is what the AI company does with the picture.
What happens to a photo of my child when I upload it?
It goes to the company's servers, and what happens next depends on your settings. Take ChatGPT as the example, because it is what the trend uses. I read OpenAI's help page on how content is used to improve models on October 19, 2026. It says: "When you use our services for individuals (Free, Plus, Pro), we may use your content to train our models." To opt out, you turn off "Improve the model for everyone" under Settings, then Data controls. The page also says: "Temporary chats are not used to improve OpenAI models while they remain temporary, even if you choose personalization." For ChatGPT Business, Enterprise and Edu accounts the default is the opposite.
Three limits are worth knowing. The page does not single out photos or other uploaded files for individual accounts, so I would assume an uploaded picture counts as content. I could not confirm from it whether anything used before you opted out can be withdrawn. And other tools have their own rules, so the answer for Gemini, Meta AI or a photo-editing app is on that tool's own privacy page, not here.
OpenAI's terms of use, effective January 1, 2026, say users must be at least 13 and that under-18s need a parent or guardian's permission. That is about a child having their own account. It is a reason for a teenager not to run this trick on their own account without you.
Then there is the wider worry. Welton's quoted concern is "who knows where these platforms are going to use these pictures." That is a worry about not knowing, and PetaPixel offers no incident to go with it. The Internet Watch Foundation and the National Crime Agency's guidance for parents, launched on July 3, 2026 in the UK, is about posting and sharing: it says AI-generated abuse imagery can be made from "everyday images of a child" taken from social media, websites or messaging apps, and it asks parents, "Am I still comfortable with how my child's images might be used?" That guidance does not address parents uploading photos to AI tools privately, and I have not found a source documenting a child's photo being misused after a private upload like this. I would treat the risk as real but unmeasured.
Will the picture still look like my child?
Not always. Gadget Review's October 6, 2026 report says "Generative AI can alter a child's appearance even when you ask it not to," citing parents' accounts of changed faces, and suggests the result is better described as an "AI-stylized portrait" than a faithful photo. That matters if the picture is going on a grandparent's mantelpiece as a record of how your child looked this year.
The Take-Back Test
I use one question before any picture of a child goes into an AI tool.
The Take-Back Test: before a picture of your child goes into any AI tool, ask whether you could take it back next year. If you cannot say yes, the picture stays on your phone.
This is my suggestion, not a research finding. It works because it does not ask you to understand every privacy policy. It asks whether you could undo the decision, and for most tools the honest answer is that you do not know.
A school photo is a keepsake. An upload is a decision you cannot see the end of.
What I would do at home
These are my suggestions, ranked from lowest risk to highest.
| Option | What leaves your phone | My view |
|---|---|---|
| Ask the photographer or school whether a smaller pack or a single digital file is available | Nothing | Worth one email before anything else, in case a smaller option exists. |
| Take your own photo and use it unedited | Nothing | A child in good light against a plain wall is a perfectly good keepsake. |
| Use a normal photo editor on your phone to crop or brighten | Depends on the app | Check where the editing happens. Many edits run on the device. |
| Upload your own photo to an AI tool for a school-portrait look | A picture of your child, to a company | Passes the Take-Back Test only if you have opted out of training or used a temporary chat, and have read what the tool keeps. |
| Remove the watermark from a proof | A copy of someone else's work | I would not. |
If you do use an AI tool, I would:
- Use your own photo of your own child, never a class photo or another family's child.
- Use your own adult account, with "Improve the model for everyone" off or in a temporary chat.
- Check the face before you print. If it no longer looks like your child, it is not a record of them.
- Keep the AI portrait out of public posts.
Today: check whether your child's school sends the photographer's order form home and whether it offers a smaller option. That takes five minutes and costs nothing.
What should I say to my child?
It is their face, so their say should grow with their age.
Under 8. You decide, and it is a good moment to show them what you are doing.
You could say: "I'm going to make a nicer picture of you on the computer. It's only for us."
8 to 12. Ask first, show them the result and let them veto it.
You could say: "I'd like to use a computer program to make your school picture. Would you be okay with that? You can look at it and say no."
13 to 15. Their say counts as much as yours, because it is their face. Treat a no as a no.
You could say: "I can get your school photo done this way or the usual way. How do you feel about a company having your picture?"
16 to 18. They may refuse outright, and they may well already know more about image tools than you do. Ask what they think.
You could say: "I'm thinking of doing it with AI to save money. Any reason I shouldn't?"
What I see in schools
The pattern I see most often is that schools treat photo day as the photographer's business and AI at home as the family's, so nobody has said anything to parents either way. A school that needs an official photo for a student ID, a yearbook or a safeguarding record may have its own requirements, and it may not accept an AI-restyled image. If you are not sure, ask the office before you spend an evening on it. If you would like to ask your school about its wider position on AI, ten questions to ask your child's school about AI is the list I would send.
What none of this can guarantee
An opt-out setting governs one account on one tool. It cannot tell you what a different app does, and it does not guarantee you can reverse something already used. A temporary chat is a promise from a company, not something you can check. The law on watermarks varies by country, and I am not a lawyer. If you are unsure whether you can use a proof, the photographer's terms and the school's rules decide that, not me. What ChatGPT keeps about your child covers the data side in more depth, and setting up ChatGPT's parental controls covers what those controls do and do not do for a teenager's own account. For the bigger picture of what to hand over to AI and what to keep, AI for busy parents is the place to start.
If your child's school has not spoken to parents about AI yet, this is the work I do with schools: helping staff write down what parents are allowed to ask. It is worth sending to your child's principal (head teacher in the UK). I also write a Sunday newsletter that keeps up with stories like this one without the panic. Parents are very welcome.
Dan Fitzpatrick is a former teacher and assistant headteacher, a Forbes contributor and the founder of The AI Educator. Read Dan's story.
