Will Ranjan-Churchill

How do you protect kids from AI?

Some thoughts on AI, and what it means to me as a parent.

A little while ago, my son turned one, and we had a party. During the party, we took a big group shot of the extended family. One family member, for various reasons, did not want to be in the group photo. However, some weeks later, they apparently regretted this. No matter, they thought, and uploading that group photo into AI, they asked a chatbot to pop them into the shot.

When I found out, I was apoplectic.

This person had taken a photo with 20 other people in it, including my son (and, by the way, a number of other children), uploaded it to a commercial AI model alongside a separate photo of themselves, and asked the AI to put them in the group photo. Not a single other person in that photo had given consent for that to happen. Some of the people in that photo cannot give consent. A good portion of people in that photo would not give consent, had they been asked (which, again, they weren't).

I felt violated. More than that, I felt - genuinely - like I had failed as a parent in protecting my son. I feel ashamed.

But the thing is, how mad can I really be? I spend most of my waking day thinking about data security, privacy, and the impact of AI (and other technologies). But others? Normal people with normal jobs?

Who is using AI?

This year's Charity Digital Skills Report tells us that nearly 80% of charities are making use of AI. But that's number of organisations, not individuals. If we zoom out, it becomes harder to get concrete stats, but here are a few:

I guess I have two points. The first is, AI isn't nearly as ubiquitous as tech people would have you believe (and I say that as a "tech person"). Secondly, even if those numbers were all above 90%, does that mean people would understand the dangers? Do most people even understand how AI works?

I travel almost every single day - either by car, train, bus, or a combination. Would I be able to swap out my car's catalytic converter if it broke? If the points system on my train home stopped working, would I be expected to roll up my sleeves, jump down onto the tracks, and have a go at fixing them? No. Obviously no.

A general understanding of tech doesn't guarantee any understanding of AI

When inviting people to our son's birthday, we'd asked them beforehand not to share photos of our kid on social media. Everyone understood, and as far as we know everyone stuck to that. But it's 2026 - we've had Facebook for 20 years. And it's taken almost all of that time for people to understand the implications and dangers involved. And not, it shouldn't be ignored, without considerable effort from activists, a lot of court cases, and very much against the wishes of very rich, powerful, and influential people. Getting the average person to understand not to post a picture of a child without permission to their Facebook wall took enormous amounts of work by a lot of people over a long period of time.

And the thing is, for some of us those lessons translate quickly and easily. For some, if you understand that posting a photo to Instagram means Meta has license to do what they want with that photo, then its a short logical leap to understanding that uploading a photo to ChatGPT means OpenAI similarly can do what they want with that photo.

But for others, the logical leap doesn't occur to them. Uncharitably, I want to call it a lack of critical thinking skills. Or, maybe, its a concerted effort on the part of tech companies everywhere to obfuscate their intentions. OpenAI are hardly going to prompt you every time you try and upload a photo:

Oh, by the way, just in case the photo you're about to upload is the photo of a child, when you upload this photo of a child, we will keep a copy of this photo of a child forever (unless you explicitly ask for its removal). We will use this photo of a child to train our models to recognise and understand the differences between this child's face and other children's faces, other adults' faces, and really anything else we chose to do with it. If another entirely unrelated person asks us at some point in the future to generate an image of a child (for whatever purpose they might want an image of a child...), your photo of a child might be used in part or in its entirety to help create that image of a child for that unknown person. That virtually created image of a child might have eyes we've pulled directly from the photo of the child you send us! Isn't that cool? Don't worry, we'll keep all your data safe, right next to all the other data we've collected about you from our interactions. We'll just hold it here, not to do anything with it. Other than build a picture of you and your life, and the people you know, so that we can better understand you as a person, people by proxy, and how best to sell our next products to you as a person, and people by proxy. So, do you still want to upload that photo of a child?

In the planning of our son's birthday, we knew from the start that we didn't want people sharing photos of him on social media. But it didn't occur to us to also explicitly ask people not to upload photos of him to ChatGPT. Why would it have?

Where does it end? Ok, uploading a group photo to ChatGPT and asking it to include you is one thing. But what if someone had taken a photo on their camera, uploaded it to Adobe Lightroom, and used any of Adobe's famously "we don't give a tiny rat's ass about your privacy" AI tools to enhance it? Not even with the intent of posting on socials. Maybe they just wanted the photo colour balanced for their own personal album, and it was easier to let the AI do it. For his second birthday, will I need to issue a 6 page T&Cs sheet of software you can and can't use if you're going to take photos? Will we need those pouches you get at gigs now to lock people's phones away? I'll have to ban Meta glasses, for sure. But now also apparently Apple watches?

So, what now?

I'm still beyond angry. But it's merged into the same background anger I feel at the state of big tech, and the way we've let our whole worlds be taken in by their greed.

And as for my shame as a parent? I don't know... I don't know yet how to navigate a world where information is so unevenly distributed, while at the same time protecting my son from the implications of that.

And I really don't know how to square away working day-to-day with tools that so easily draw out such a visceral reaction from me when they aren't used the way I think they should be.

ai, digital privacy, parenting


Want to chat about anything in this post? I'd love to hear from you!

Feel free to drop me an email

Or find me on Mastodon


⬅ Previous post
Weeknote #37

Next post ➡
Weeknote #38