Over the past year, artificial intelligence has become increasingly better at generating photorealistic images of people.

At the same time, a subculture focused on explicit and pornographic content has emerged – material that the major AI tools have banned. Around the turn of the year, a group with the ambition to improve and simplify the production of AI pornography raised over half a million kronor on Kickstarter in just a little more than a week, before the collection was stopped.

And in the dark shadow of this development, another variant has emerged: AI-generated images depicting child sexual abuse.

"This type of image poses a significant challenge"

For SVT News, Europol confirms that this type of material already exists – and is spreading.

"Europol is aware of the existence of this type of AI-generated material in the EU. Advances in AI have made it easier for criminals to create child sexual exploitation material, writes Claire Georges, spokesperson for Europol.

"AI technology that can generate these types of images poses a significant challenge.

The Swedish police share Europol's view of the situation.

Parallels to the "Manga Case"

But would the creation of this type of AI images be illegal in Sweden?

In the high-profile "Manga" case, a translator specializing in Japanese cartoons was charged with child pornography offenses for possession of cartoon images. The translator was acquitted in the Supreme Court.

The Supreme Court found that the cartoons were pornographic and depicted children, but that they were fantasy characters that cannot be mistaken for real children.

According to the ruling, "Criminalizing the possession of the cartoons would go beyond what is necessary."

Would this also apply to AI-generated images?

In the video above, Björn Sellström, Commissioner at the National Operations Department (NOA) at the police, explains the legal situation in Sweden – and what problems this AI development can bring.