Facts
Through social media, the defendant asked an unidentified person to produce composite images combining a 16-year-old's face photo with nude photos of an unidentified woman, and received them. He was charged with producing child and youth sexual exploitation material on five occasions under the Act on the Protection of Children and Youth against Sex Offences, which punishes production with life imprisonment or at least five years. The appellate court acquitted him on that count, finding it difficult to treat the composite images as sexual exploitation material within the meaning of the Act. The classification question then came before the Supreme Court.
Question
Article 2(5) of the Act divides exploitation material into two categories: material in which a child or youth appears, and material in which a person or representation clearly recognisable as a child or youth appears. The Supreme Court had to decide which category covers composite photos or AI-generated deepfake videos that put a real minor's face onto sexual imagery, and what test determines whether they fall into the second. The question went directly to how existing criminal provisions apply to sexual imagery generated with AI tools.
Decision
On 14 August 2025 the Supreme Court held that the first category covers material in which a real child or youth actually appears and engages in sexual acts. Composite photos and deepfakes built on a real minor's face, including those edited with AI, are images created by their maker and fall outside it. Where their content depicts sexual acts by a child or youth, however, they can be representations clearly recognisable as a child or youth. This is judged on all the information, such as the person's appearance and physical development, actual age and identity, the image's source and production, and its background and setting. Applying these principles, the Court accepted the appellate court's conclusion and dismissed the appeal.
Why it matters
Published by the Court as one of its important judgments, the decision clarified which legal category applies to AI imagery built on real minors' faces. The first category is confined to images in which a real minor actually takes part, while deepfakes are assessed under the representation category through a multi-factor examination of content and context. The ruling supplies the test for deciding when such imagery is punishable as exploitation material.