In 2023, tech companies sent roughly 4,700 reports of AI-related child sexual abuse material to a US clearinghouse. In 2025, they sent more than 400,000. That is an 8,400% increase in two years, and the people who study this for a living are begging governments to treat it like the emergency it is.

The Number That Should Wreck Your Afternoon

Here is the headline stat again, because it deserves a second look. According to research from the Edinburgh-based Childlight Global Child Safety Institute, which examined reports that tech companies filed with the US-based National Center for Missing and Exploited Children, AI-related abuse material went from about 4,700 reports in 2023 to more than 400,000 in 2025.

That includes deepfakes made with so-called nudify apps. Yes, those exist. Yes, they are apparently easy enough to use that the volume of reports multiplied nearly ninety-fold in the time it takes a toddler to learn to read.

The images are fake. The harm is not. Researchers say they are used to blackmail children into handing over money or more images. A fabricated picture becomes a very real weapon the moment someone threatens to send it to a kid's parents, classmates, and coach.

A Second Watchdog Is Seeing the Same Fire

Childlight is not alone. The Internet Watch Foundation, a UK-based NGO that monitors abuse material globally, reported last month that AI-generated material is still growing online. Its analysts found more photorealistic child sexual abuse material in the first half of 2026 than in all of 2025.

The specifics: the IWF assessed 6,310 AI images that met the legal definition of child sexual abuse this year. Last year's total was a little over 4,500. That is a 40% jump, and the year is not even over yet.

Two separate organizations, using different methods, landing on the same conclusion. When the smoke detectors in two different rooms go off at once, you do not get to call it a faulty battery.

Yes, Better Reporting Explains Some of This. Not Most of It.

Let's be fair, because the researchers were. Prof Debi Fry, Childlight's global director of data, acknowledged that improved detection and reporting may explain some of the increase. More companies looking harder means more reports filed.

But nobody serious believes better paperwork turns 4,700 into 400,000. Fry called the speed of the change "deeply concerning" and said it underlines why governments and tech companies must act urgently to understand these emerging forms of abuse and get the material removed quickly.

Improved detection explains why we are finding more. It does not explain why there is so much more to find. Those are different problems, and the second one is the one that should keep people up at night.

The Offline Numbers Are Not Exactly Comforting

The AI surge is the headline, but the broader findings are grim in their own right. Childlight estimates that more than 300 million children worldwide are victims of at least one form of technology-facilitated sexual exploitation or abuse every year. It also estimates that about one in four children will be subjected to at least one form of abuse before turning 18.

In North America, almost a third of girls and more than one in 10 boys experienced in-person sexual abuse during childhood, according to the research. That covers sexual assault, rape, and upskirting.

And the trend line in the US is moving the wrong way fast. The recorded rate of reports and notices involving child sexual abuse material rose 63% between 2024 and 2025. Canada's rose 18%. Same continent, wildly different trajectories.

What the Researchers Are Actually Asking For

Childlight's chief executive, Paul Stanfield, said governments are too often working with an incomplete picture of where harm is happening and who is most at risk. You cannot fix what you refuse to measure.

The ask is not exotic. Recognize child sexual exploitation as a preventable public health emergency. Invest in better evidence. Turn that evidence into action. Stanfield's point is that there is nothing inevitable about any of this, and that the knowledge and tools to make a difference already exist.

That framing matters. We treat other preventable harms with surveillance systems, funding, and coordinated response. Children being sexually exploited at industrial scale by software gets a strongly worded report and a news cycle.

The Dingo Take

We built the machine, shipped it, and are now surprised it does what machines do.

The nudify apps did not appear by accident. Somebody wrote them, somebody hosted them, somebody processed payments for them, and somebody is making money. Meanwhile the tech industry that spent three years promising us that AI would cure cancer and write our emails has been remarkably quiet about the 400,000 reports of kids being exploited with it. Where is the keynote slide for that number?

Governments are not off the hook either. The researchers are begging for basic data collection and a public health framing, which is the political equivalent of asking someone to please look at the burning building. If lawmakers can find the time to fight about everything else, they can find the time to make it expensive and illegal to build, host, and profit from the tools that do this. Every month they spend studying it is another few tens of thousands of reports. The curve is not waiting for a committee.

Sources