Does it infringe copyright to make an unauthorized copy of artwork and to use that as a base image for the creation of a commercial product that is substantially similar to the original work using an AI tool to do minor visual editing to produce the commercial product? Yes, according to a recent decision of the Beijing Tongzhou Court.
In the case, individuals were convicted of criminal copyright infringement when they used an AI tool, described as a “picture-to-picture’ method to generate effects”, to copy other peoples works of art making only minor modifications to create puzzle pieces for a jigsaw puzzle commercial product. In May 2025 five defendants were convicted. Here, the gravamen of the complaint was not content produced by an AI system, but rather infringement in the inputs that were reproduced in outputs with only minor visual effects changed. According to a report on the case,[i] the judgment of the court included the following reasons for decision:
According to the relevant provisions of the Criminal Law of the People’s Republic of China and the Copyright Law of the People’s Republic of China, to determine whether the subject matter of the case constitutes a work, it must meet the following conditions: first, it must be original; second, it must be a creation in the fields of literature, art and science; third, it must be able to be expressed in a certain form; and fourth, it must be an intellectual achievement. Therefore, “originality” is a necessary condition for a work. The key to examining whether a work is “original” lies in whether the work has “minimum intellectual creativity”, that is, whether the author has invested intellectual labor in the creative process.
In this case, the defendant used generative artificial intelligence technology to generate pictures and converted other people’s works of art into commercial puzzles. The generation of puzzle pictures only uses the preset parameters of generative artificial intelligence, that is, the “picture-to-picture” method to generate effects. The defendant did not invest intellectual labor in the generated puzzle pictures, such as setting the number of iteration steps, modifying the random number seed, adding prompt word content, etc. The degree of creative control and contribution to the generated results is minimal. Therefore, although there is a very small proportion of inconsistency between the puzzle picture and the original work of art, it should still be identified as a picture that is substantially the same as other people’s works of art, and the defendant’s actions should also be identified as copyright infringement…
[i] Beijing Law Network, June 27, 2024, Ji Xinrui, @ https://mp.weixin.qq.com/s/xaUSThKXwaOE6yq20Qbj1g (translated by Google).
