Artificial intelligence is misused to create sexualized images of children – news Trøndelag

– This type of abuse material has almost no boundaries, says Svein Øverland. He is a specialist in psychology and works to help people who have pedophile thoughts. Now he fears the consequences of AI-generated images of abuse being spread. – Such images can increase the risk that a person with pedophilia will abuse a child. According to Øverland, the fact that the images are not real can give those who look at them less of a guilty conscience. He does so worriedly. – There are many pedophiles who never molest children, but the more someone masturbates to children, the stronger their drive will become. Øverland is concerned that the threshold is lowered for using images of abuse when they are made with AI. Both because they become more accessible, and because of the fact that they are not real can ease the conscience. Photo: Erlend Lånke Solbu / news – Borderless crime British media recently revealed that pedophiles use artificial intelligence to create realistic abuse material. The BBC writes. According to Kripos, this is only becoming more and more common in Norway as well. Section leader Katrine Olsen says that Kripos attaches great importance to building expertise in the field of AI as the technology is constantly becoming more advanced. Photo: PRIVAT Now they are preparing for a large increase, says Katrine Olsen, section leader in the section for internet-related assaults in Kripos. Olsen also says that they have monitored several abuse forums where it is expressed that they want more AI-generated images. – This is a borderless crime and in many countries this will not be prosecuted. She says that different practices in different countries make the fight against abuse material problematic. In Norway, this falls under Section 311 of the Criminal Code on depiction of sexual abuse against children or depiction that sexualises children. What are AI-generated images? Images created by artificial intelligence. Based on the text you enter, the software creates images based on a database with huge amounts of different images and concepts. Three of the most used services for AI-generated images: DALL-E, MidJourney and Stable Diffusion. Example: If you ask artificial intelligence to create a pink horse on the moon, it will do it because it has seen pictures of horses, pink and the moon. The result will be a composition of the various concepts. Source: Public Broadcasting Service (PBS) & Forskning.no Lack of regulations for artificial intelligence – The images are photorealistic, so distinguishing between AI-generated and real abuse material could potentially become completely impossible, says Inga Strümke. She is an associate professor at NTNU and has written a book about artificial intelligence. She fears AI-generated images will make it much more difficult to trace actual victims. – That this type of technology is misused is not surprising. When you develop powerful technology, it is almost guaranteed that it will end up being misused. She believes that a KI Act on regulation cannot come quickly enough. – The Internet has no national borders. Norwegian users can easily find such images even if they are on foreign sites, says KI researcher Inga Strümke. Since 2021, the EU has been working on an AI law, the EU’s so-called “AI ACT.” It was recently voted on but has not entered into force. Inga Strümke is an associate professor at NTNU and has written a book about artificial intelligence. Photo: Ine Julia Rojahn Schwebs



ttn-69