Japan Alleges AI Image Copyright Violation in Landmark Case

Japanese authorities have charged an individual with the unauthorized reproduction of an AI-generated image. This marks the country's first legal case where an AI-created image is being treated as a copyrighted work under the Copyright Act.

As reported by the Yomiuri Shimbun, a man in his 20s from Chiba prefecture used Stable Diffusion in 2024 to create the image in question. Police allege a 27-year-old man, also from Chiba, then used this image without permission for the cover of a commercially sold book.

The image's creator informed the Yomiuri Shimbun that he used more than 20,000 prompts to achieve the final result. Investigators concluded the creator played a sufficiently active role in the image's generation. The case has been forwarded to the Chiba District Public Prosecutors Office.

Japan's Copyright Act defines a protected work as a "creatively produced expression of thoughts or sentiments" in literary, academic, artistic, or musical fields. The Agency for Cultural Affairs has clarified that an AI image created with minimal or no human instruction does not meet this definition and thus cannot be copyrighted.

Conversely, if an individual uses AI as a tool to creatively express their ideas or emotions, the resulting output may qualify for copyright protection. This determination is made on a case-by-case basis, evaluating the creative process. Key factors include the detail of the prompts, iterative refinement of instructions, and creative selection or editing of the outputs.

As a precedent-setting case, legal experts are closely watching. An attorney specializing in AI and copyright from the Fukui Bar Association told the Yomiuri Shimbun that sufficiently detailed and specific prompts could lead to an AI output being recognized as a copyrighted work. The final image is assessed based on how well it aligns with the creator's pre-existing vision, which is evidenced by the specificity and iterative refinement of the prompts. "The crucial point is determining whether the person used prompts to achieve a specific, anticipated result," the lawyer stated.

In a related trend, AI-generated content that closely mimics existing copyrighted works has sparked significant controversy in Japan. This is particularly true for Sora 2-generated videos featuring famous Japanese characters. In response, the Japanese government and an organization representing major creators like Bandai Namco, Studio Ghibli, and Square Enix have demanded that OpenAI stop unauthorized training on Japanese intellectual property.