AI-generated images of child sexual abuse are increasing sharply. The Internet Watch Foundation found 6,310 of them in the first half of this year, 40 percent more than in all of 2025, and Dutch experts told NU.nl Tech they expect the number in the Netherlands to rise as well.
What happened
NU.nl Tech reports new figures from the Internet Watch Foundation (IWF), which searches for and reports child sexual abuse material online. In the first six months of the year the IWF found 6,310 AI-generated images showing the sexual abuse of children. That is 40 percent more than it recorded in the whole of 2025.
For 6,221 of these images, the IWF could establish the sex and age of the child shown. In 98 percent of those cases the images showed girls. The largest groups were children aged seven to ten (2,534 images) and children aged eleven to thirteen (2,369 images). The IWF also found 190 images of babies and toddlers younger than two.
Dutch experts were not surprised. Offlimits, the Dutch expertise centre on online abuse, saw the number of AI abuse images almost double last year compared with 2024. Its director, Robbert Hoving, told NU.nl that the real difficulty is telling fake images from real ones. Early AI images gave themselves away through obvious mistakes such as extra fingers. Now, he said, the main clue is often that a photo looks too perfect. That causes a practical problem. If Offlimits sends police an image it believes is real and it turns out to be AI-generated, the police may end up searching for a victim who does not exist.
On the law, lawyer Charlotte Meindersma told NU.nl that AI-generated images of child abuse are always illegal. Under Dutch law it makes no difference whether the child in the image exists. The aim, she explained, is to stop this material from becoming normal. People who make it face up to four years in prison or a fine of up to 110,000 euros.
Hoving argued that even images of invented children can involve real victims. An AI model does not come up with what a child looks like by itself. It learns from existing images, so its training data probably includes images of real victims.
Enforcement is where the system falls short. Arda Gerkens of the Autoriteit online Terroristisch en Kinderpornografisch Materiaal (ATKM), the Dutch regulator for online terrorist content and child sexual abuse material, told NU.nl that her organisation can do only a limited amount, working with other regulators, the police and the Public Prosecution Service. Banning the AI tools outright is not simple, she said, because they can also be used for legal purposes. She compared them to a phone camera, which can be misused but is not banned for that reason. Gerkens called it a complicated situation with no ready-made answer, but warned that the problem will escalate if the tools are not tackled. NU.nl reports that the ATKM is now looking into a new way to get abusive tools taken offline.
Why it matters for internationals in the Netherlands
The tools are ordinary consumer technology
The most important point in the NU.nl report is that the technology behind this abuse is not hidden or specialised. Gerkens’s comparison to a phone camera shows how she sees the problem: these are general-purpose tools that are widely available and can be used for good or harm. NU.nl reports that despite repeated warnings from groups such as Offlimits, these images remain easy to make. For families, that means the risk does not come only from the dark corners of the internet. It sits close to the apps and devices people already use.
Real children, invented situations
NU.nl notes that AI images can show real, identifiable children in situations that never happened. The IWF’s position, as quoted by NU.nl, is that AI-generated abuse material causes real harm whether or not the image was edited. Parents who share photos of their children online should know that, according to the experts NU.nl spoke to, a real child’s face can become the starting point for this kind of abuse.
Schools and young offenders
For international families with children in Dutch or international schools, the most direct finding is about who makes these images. Gerkens told NU.nl that a large share of the people making AI abuse images are minors themselves. They often treat it as a joke, she said, but it is not one, and the images sometimes spread through an entire school. Offlimits reports that staff on its prevention line increasingly have to explain to callers that making these images is illegal.
Parents coming from countries with different laws should know where Dutch law stands. As Meindersma explained, it does not matter whether the child in the image exists. Making such an image is illegal either way. Hoving also raised a wider concern: children who see extreme content online too early can become desensitised, and some then look for more extreme material. He said this does not apply to all children, but it is one reason Offlimits believes children should not be able to visit certain websites.
EU AI rules and the Dutch data protection authority
The report also points to a change in oversight. From 2 December, the Autoriteit Persoonsgegevens, the Dutch data protection authority, will supervise compliance with the European AI Act. Hoving told NU.nl he hopes the regulator will enforce strictly against this kind of abuse material. For internationals who already know the Autoriteit Persoonsgegevens from privacy and data protection matters, this means the same regulator will soon also oversee AI systems in the Netherlands.
What to watch next
The first date to watch is 2 December, when the Autoriteit Persoonsgegevens starts supervising the EU AI Act. How it acts against AI tools used to make abuse images will show whether Hoving’s hope for strict enforcement is met.
The second is the ATKM’s work on a new way to get abusive AI tools removed. Gerkens has said that without action against the tools the problem will escalate, so whatever the ATKM decides will show whether Dutch regulators can go beyond removing individual images.
The third is the Dutch figures. Offlimits saw AI abuse images almost double last year, and Dutch experts expect the numbers to keep rising. Further data from Offlimits and the IWF will show whether the international trend is repeating in the Netherlands, and whether awareness efforts aimed at young people, including in schools, are having any effect.
Source: NU.nl Tech
Sources
- NU.nl Tech: AI-kindermisbruik groot probleem: 'Beelden gaan soms hele school door', accessed 8 October 2026
The Netherlands, for internationals
One email a week: the week's news for internationals in the Netherlands, new practical guides and what changed in the cities. No spam, unsubscribe any time.
