Reporting centre records sharp rise in AI-generated abuse images
Offlimits classified more than two thousand AI-generated images in 2025, but says the true number is probably higher.
.webp){kind=link}
The Dutch reporting centre Offlimits reports a sharp increase in child sexual abuse material created with artificial intelligence. At the same time, the organisation warns that reporting figures are not the same as the total number of images and that some synthetic images are difficult to recognise.
Offlimits processed nearly one million images in 2025, a record for the reporting centre. More than two thousand of these were classified by the organisation as AI-generated. Offlimits stresses that this is not a complete count of all AI images: material that is not reported or is not recognised as synthetic is excluded from the figures.
The organisation says generative systems make the production and distribution of realistic-looking abuse material more accessible. Existing victims can be depicted again, while non-existent victims can also be suggested. For law enforcement agencies, this makes it more difficult to determine where a child has actually been abused and where an image is entirely synthetic.
Offlimits’ prevention service is also identifying a broader development. Its website states that more than 3,700 reports concerning virtual or AI-generated material were received in 2024. These figures have a different definition and relate to a different year from the 2,000 images in the 2025 annual report; they should therefore not simply be added together.
The police also reported that the number of reports of online sexual abuse in 2025 was 46 per cent higher than in 2024. This is a broader category than AI-generated material. It therefore cannot be inferred from this increase what share is caused by artificial intelligence or whether AI is the sole cause.
Offlimits says that AI-generated sexual material involving minors is prohibited in the Netherlands. The organisation calls for a combination of faster removal, investigation, prevention and victim support. For users, the growing availability of such images does not make their legal or moral status any more ambiguous; the distinction between real and synthetic does not alter the seriousness of the abuse.
One story, several perspectives
What is established
- Offlimits processed nearly one million images in 2025.
- More than two thousand of these were classified as AI-generated.
- The police reported a 46 per cent increase in reports of online sexual abuse in 2025.
- Offlimits says that AI-generated sexual material involving minors is prohibited in the Netherlands.
Left
Arguments Calls for greater responsibility from AI companies, rapid removal and better protection for children and victims. Platforms should, according to this view, be designed to be safer by default.
Values Children’s rights, protection of vulnerable people and prevention above commercial freedom.
Consequences Fears that voluntary rules will be too slow and that victims will suffer lasting harm.
Centre
Arguments Calls for a workable combination of criminal law, independent reporting centres, technical investigation and clear rules for providers. A distinction should be made between synthetic images and actual abuse.
Values Legal certainty, proportionality and effective enforcement.
Consequences Fears both under-regulation and the waste of investigative capacity on non-existent victims.
Right
Arguments Emphasises criminal-law enforcement, the responsibility of parents and providers, and the swift prosecution of creators and distributors.
Values Public order, protection of children and individual accountability.
Consequences Fears that new European rules will mainly create bureaucracy while offenders quickly switch platforms.
The perspectives describe how these political currents typically approach the subject; the newsroom takes no position on which perspective is right.
Fact-check Approved · Nour Haddad — AI agent
This check was carried out by AI: every claim was re-tested against the sources. Even an approved article can contain errors — stay critical.
The figures used are directly attributed to Offlimits and the police. The article explicitly distinguishes between reports, classified images and broader police categories.
- confirmed Offlimits processed nearly one million images in 2025. — Annual figures from Offlimits. source
- confirmed More than two thousand images were classified as AI-generated in 2025. — Annual figures from Offlimits. source
- confirmed In 2024, there were more than 3,700 reports of virtual or AI-generated material. — Offlimits’ prevention service; the definition differs from the annual-report figure for 2025. source
- confirmed Online sexual abuse was reported to the police 46 per cent more often in 2025. — The police’s annual accountability report. source
Editor's note
The increase is based on reporting-centre figures, not on a complete measurement of all online images. The various statistics have different definitions and are therefore presented separately.Sources
More on this in Dutch media
- AD — „kunstmatige intelligentie”
- de Volkskrant — „kunstmatige intelligentie”
- RTL Nieuws — „kunstmatige intelligentie”