In 2025 the illegal videos produced by AI exploded: +1,286 in the first six months. Criminals exploit accessible models to generate increasingly realistic content, spent based on real victims. Here’s what emerged from the new data collected by the Internet Watch Foundation.
Artificial intelligence is also used for generate child pornographic videos. A new frontier of online abuse is taking shape, and according to the data of theInternet Watch Foundation (IWF) The phenomenon is growing worryingly. In the first six months of 2025, the IWF identified 1,286 illegal videos artificially generated. Last year, in the same period, there were just two.
Most of these videos – over 1,000 – is part of Category Athe most serious, and many are based on images of real victims. Creating these clips is very easy, just train a model with a few real content to generate an infinite number of fictitious but convincing videos. And in fact, as the IWF explained these contents are now almost indistinguishable from real material. “Those who abuse have a one available today powerful technology, easily accessible, and constantly improving. And he is taking advantage of it. “
The invasion of the web: +400% content a to the first months of 2025
Access and simpler access to Advanced video generation modelsallows criminals to create ultra -ruling child pornographic content. Some users, explained the IWF, They train the models to the real videos of abuse On minors, some clips are sufficient to obtain extremely realistic results. “Sector AI is highly competitive and very rich in investments: unfortunately, this provides the bad guys a wide range of tools to be exploited “, The IWF explained.
Artificial intelligence arrives in the Chamber: deputies will have a chatbot to write the laws
In the first six months of 2025, the IWF recorded an increase in 400% of the URLs containing child pornographic content generated with AI: 210 reports against 42 of 2024. Each web page often hosts hundreds of images and videos.
Even if the contents are “fake”, The victims can be real: Many videos to the real images of children extracted from social media, family albums online, or even webcam hacked. IWF analysts explained that most of the images seem to have been created with basic artificial intelligence models available free Online. On Forum del dark webpedophiles openly discuss the rapid evolution of these tools, share advice on available software, which prompt to choose and on Best models to use.
New frontiers of online crime: what are the risks
To counteract this new threat, the United Kingdom has introduced more severe laws. It’s time illegal to own, create or distribute artificial intelligence tools designed to produce sexual abuse content on minors. Those who transgress risks up to five years in prison.
In addition, i are also prohibited manuals that teach to create these contents with the AI: possessing them or sharing them can involve up to three years in prison. The interior minister Yvette Cooper underlined the importance of face the phenomenon both online and offlineexplaining that the protection of minors requires tools and laws adequate to the new digital challenges posed by artificial intelligence.
According to Derek Ray-Hill, interim CEO of the IWF, if you do not intervene quickly, sexual abuse on minors through ai risks saturate the web with childhood content. Ray-Hill warns: “We are faced with a possible explosion out of control of child pornographic content generated by AI, with devastating impacts not only online but also in the real world: trafficking of minors, slavery and abuse could intensify“.”
