CaraComp
CARACOMP DEEP DIVE

AI Fakes Explained: 150 Photos, 49% Of Schools Hit

September 13, 20263:54
AI Fakes Explained: 150 Photos, 49% Of Schools Hit
Chapter 1 of 1 · Cold open
0:00 / 3:54

Full Transcript

One school got an email like a nightmare. Attached were one hundred and fifty fake photos of its own students. Real faces, a fake scene, and a gang wanting money. One clear photo of your child is all it takes. By the end, you'll know what to do the night it happens. And the one check that clears it in minutes.

The UK's watchdog checked every one of them. All counted as illegal, under UK law. That's how serious this already is. Hold that thought.

They didn't hack anything. They didn't need to. One class photo was already public. That's all a fake needs to start.

Here's why it matters more. Forty nine percent of UK schools already had a photo like this. That's about one in every two schools. Nearly half of them. Not some far off risk.

You post that same school photo. A smiling face, a school trip, nothing more. Somewhere, a stranger opens it and saves it. That's the moment it stops being yours.

That's the gap criminals use. And one school just found out how far it goes.

Late last year, an email lands at one UK school. Attached: one hundred fifty photos of its own students. Real faces, dropped onto fake, illegal scenes. Then the demand: pay up, or they go public. A safety group finds every last one.

Malwarebytes talked to safety experts about it. The UK's safeguarding minister called it exactly what it is. A deeply worrying emerging threat.

Here's how a fake gets made, in minutes. One clear photo. One AI tool. One fake scene, and it's done.

Right now, someone still does this by hand. Experts warn it could turn automatic. Then it stops being rare.

If this helps, a like and a subscribe help us keep it coming.

So if it turns automatic, it won't stay rare long. It's already hitting schools everywhere. Here's how big the gap really is.

So how big is the gap? Ninety three percent of head teachers already know AI can do this. Only forty three percent feel ready to respond. That's fewer than one in two. That gap is the real danger.

And the numbers are climbing fast. Reports of AI made abuse images more than doubled in one year. From one hundred ninety nine, to four hundred twenty six.

This isn't one bad apple. It's happening in ordinary classrooms. Any photo, any school, could be next.

So what actually protects you here? One tool is spreading fast. It's called facial comparison. It checks if two photos show the same real face.

This is where the story started. The watchdog fingerprinted every image. That's how platforms block it for good.

Here's the part everyone gets wrong. A face match is one lock. It proves whose face it is. It does not prove the scene is real.

So one lock was never enough. You need the second one too. Together, the two locks tell the whole truth. That second lock is what actually protects you.

So how does this actually work? First, upload a real reference photo you trust. Then compare it to the photo in question. You see a percentage, in minutes, not days.

We wrote up the story ourselves. In plain words, it asks one question. Same face, or not?

This is what checking it actually looks like. A printed photo, next to a screen. Side by side, face by face.

This is the report behind that gap. Ninety three percent knew. Only forty three percent felt ready. Same numbers, straight from the page.

Here's what the second check actually adds. Facial comparison confirms identity. It doesn't confirm the scene actually happened. It adds the scene, the edits, and where it came from.

So here's what to do, before you panic. Report it, and never share it further. That stops it from spreading. Save it, and write down the date. That gives you proof. Check it against a trusted, real photo. That tells you for sure.

Reporting already works. Childline's own reports jumped by a third last year. That is real proof it helps. Every report adds to a system built to catch these fast.

Skip the second check, and a fake passes for real. That costs a family the time to stop it.

Here's what that second check looks like in practice. One face, checked against three real photos. Each one gets its own score.

So what do you do tonight? Ask for the check that covers both. Matched face, not matched scene. Ask for both, every time.

Every day we take one story like this and make it simple. Subscribe and tap the bell. Tomorrow's is free too. See you in today's briefing.