Shanghai Court Rules on the First AI Prompt Related Copyright Case in Shanghai

On November 6, 2025, the People's Court of Huangpu District, Shanghai (“the Court”) issued its first-instance judgment regarding a case involving AI prompt copyright infringement, which is the first case of such type in Shanghai, and rejected all claims brought by the plaintiff. 

According to the relevant report, the plaintiff is a company engaged in artistic creation and has utilized the Midjourney platform to generate images for commercial use since 2022. Its creative process involves writing and inputting a set of prompts, after which the AI platform generates four images simultaneously. Then, the plaintiff selects one image and requests the AI to refine specific details until an image meeting its requirements is produced. It is worth noting that during the courtroom demonstration of the image generation process using this AI platform, the AI platform was unable to produce exactly the same artwork even when the user employed the identical prompt text.

In this case, the plaintiff created and used six sets of prompt phrases in 2022 to generate artwork on the said AI painting platform, and published the works on platforms such as Xiaohongshu. The content of these prompts encompassed the artwork's artistic style, primary elements, and material details, specifically including: “Art Nouveau style illustration of Aquamarines Stygiomedusa gigantea,” “by Alphonse Maria Mucha,” “Ancient hand-painted manuscripts,” “Papyrus,” “Complex and delicate jellyfish texture,” “Gorgeous gold inlaid wooden picture frame,” and “Mirror symmetry.”

It is reported that since the said AI painting platform automatically publishes users' previously used prompts, and the plaintiff habitually searches for their own works using keywords, they discovered during one such search that the two defendants in this case were using identical prompts written by the plaintiff to generate images. These images bore a close resemblance to the works generated by the plaintiff. Accordingly, the plaintiff filed a lawsuit with the Court, asserting that “the defendants' act of generating and publishing artwork on the Midjourney platform using its created prompts infringed upon the plaintiff's rights to reproduction, distribution, online dissemination, and attribution for the literary work. Therefore, the plaintiff requested the People's Court to order the defendants to cease the infringement and compensate for reasonable expenses incurred in protecting the rights amounting to 9,900 yuan.”

The defendants replied that the prompt words involved herein constitute ideas, being merely a combination of several words and thus limited expressions. Therefore, such prompts do not qualify as works protected under the Copyright Law, and the plaintiff's act of writing them does not constitute a creative act. The defendants further argued that under Midjourney's terms of service, the plaintiff had waived all rights to the prompts, meaning that it had entered the public domain by default. Accordingly, the defendants asserted that its use of the prompts constituted “an individual's learning activity involving prompts within an open community, falling within the scope of fair use as defined by the Copyright Law.”

Upon trial, the Court considered that “the six sets of prompts involved herein employ a basic structure comprising artistic style, primary elements, materials and details, scientific context, and main composition. Essentially, they function as instructions or descriptions input by users into AI systems to guide the generation of specific images. In terms of the format, while they incorporate multiple elements, these are merely listed without any logical grammatical connection; the keyword phrases are randomly assembled, lacking both hierarchical progression and a narrative sequence within the scene. Regarding the originality, these prompts lack the author's distinctive characteristics. The selected artistic styles, material details, and other elements represent conventional expressions within the field, failing to demonstrate the author's unique aesthetic perspective or artistic judgment. Furthermore, the involved prompts merely embody abstract creative ideas and a collection of instructions, primarily listing and describing visual elements, artistic styles, and presentation forms. Such content falls more into the realm of abstract creative conception and belongs to the category of ideas. Therefore, while the prompt terms reflect certain creative intentions, they do not demonstrate the author's personalized intellectual investment at the expressive level and should not be recognized as works. The plaintiff does not hold copyright over the prompt terms and consequently has no right to assert copyright infringement.” In a word, the court dismissed all of the plaintiff's claims in its first-instance judgment.

This case concerns the determination of whether AI prompts can constitute works under the Copyright Law. Article 3 of the Copyright Law of the People's Republic of China stipulates that “The term ‘work’ as used in this Law refers to intellectual creations in the fields of literature, art, and science that possess originality and can be expressed in a tangible form, including: (1) literary works; (2) oral works; (3) musical, dramatic, folk art, dance, and acrobatic art works; (4) fine art and architectural works; (5) photographic works; (6) audiovisual works; (7) graphic works and model works such as engineering design drawings, product design drawings, maps, and diagrams; (8) computer software; (9) other intellectual creations that meet the characteristics of works.”

In the current case, the Court held that the prompt words at issue (such as “Ancient hand-painted manuscripts”) merely constituted operational instructions issued by users to the AI, which was a list of key elements, styles, and materials for the image and remained at the stage of abstract conception. They did not meet the threshold of “intellectual creations capable of being expressed in a tangible form” as required by the Copyright Law. Therefore, the AI prompt words at issue still fell within the category of “ideas” and did not constitute a work. As to the originality, while some prompts contained multiple keyword groups (e.g., “Art Nouveau style illustration of Aquamarines Stygiomedusa gigantea”), they amounted to nothing more than a “simple enumeration of elements lacking grammatical logic or narrative structure.” Moreover, the selected artistic styles and detailed descriptions are conventional expressions within the industry, failing to demonstrate the author's unique aesthetic judgment or personalized intellectual input. Consequently, they do not satisfy the requirement of originality.

In conclusion, this judgment determined that simple AI prompts do not constitute works in accordance with the definition of “works” in the Copyright Law. Excluding such prompts which are typically brief and highly functional from the scope of copyright protection not only reduces the costs for users in subsequent creation and dissemination but also fosters a more open community ecosystem. As for whether complex, narrative-driven, and highly personalized “long prompt texts” emerging in the future could constitute “works” under the Copyright Law, this will require case-by-case analysis.