Tag: intellectual property-trademarks and copyright,digital economy,artificial intelligence
Introduction
On March 19, 2025, the Chinese Zhangjiagang Court ruled in a recent AIGC copyright infringement case Feng v. Dongshan Company that, the plaintiff's AI-generated pictures lacked enough original authorship to be copyrightable and that the prompts were not copyrightable either.[1] Unlike the previous AIGC copyrightability cases where the local Chinese courts recognized the original authorship in the AI-generated work, this is the first Chinese case under which AI-generated pictures were denied copyright protection.
Background
In this case, Ms. Feng ("Plaintiff/Feng") is a designer, and she created several AI-generated pictures using Midjourney, a text-to-picture AI service, in early August 2023. She then on August 15, 2023, published the AI-generated pictures titled "Wings of Fantasy Artistic Transparent Chair" on the social media platform, Little Red Book. In her post, Feng asked if there were furniture manufacturers who might be interested in her designs. She also shared the prompts that she put into Midjourney in this post: "Children's chair with jelly texture, shape of cute pink butterfly, glass texture, light background."
Pic 1: "Wings of Fantasy Artistic Transparent Chair" created by Plaintiff
On August 20, 2023, Mr. Zhu ("Defendant A") approached Plaintiff asking whether Plaintiff would grant him a license to use the AI-generated pictures to mass-produce chairs through the Kuashi Plastic Manufacturing Company ("Defendant B"). Plaintiff refused this proposal because she was negotiating a separate license agreement with a third party at that time. However, in January 2024, Defendant A made several posts on Little Red Book promoting certain children's chair sold by the Dongshan Company ("Defendant C") bearing a trademark of 
owned by the Shenzhou Company ("Defendant D").
Pic 2: Pictures of chairs sold in Defendant C's online store
Pic 3: The actual product made and sold by Defendants
Pic 4: The packaging of the alleged infringing product
Plaintiff claimed that the chairs manufactured by Defendants B and D and sold by Defendants A and C are substantially similar to her original design, and thus brought a copyright infringement and unfair competition litigation against all the four defendants, demanding them to stop manufacturing and selling the infringing products and to compensate Plaintiff for her losses.
In their defence, Defendants argued that Plaintiff’s AI-generated pictures should not be protected under Chinese Copyright Law because (1) the prompts input by Plaintiff are merely unprotectable ideas; (2) Plaintiff had no control to the AI generative process or the final AIGC output; and (3) the expressive elements in Plaintiff's AI-generated pictures (e.g. cute butterflies) could be easily found in other pre-existing pictures posted on the Little Red Book (Pic 5). Therefore, Plaintiff's pictures should not be protected by copyright.
Meanwhile, Defendant A argued that although he used the same prompts as shared by Plaintiff in her post, he independently created the AI-generated pictures (Pic 6) through Midjourney, and then commissioned a third-party to finalize the product design (Pic 7) through multiple rounds of edits and changes. Defendant A also argued that, since both Plaintiff and himself used the same AI service app, with practically the same prompts to generate pictures, the similarity between Plaintiff's AI-generated pictures and those of Defendant A was caused by the AI's algorithm (not by Plaintiff’s originality).
Pic 5: Other pre-existing pictures of similar styles on the Little Red Book
Pic 6: Pictures generated by Defendant A using Midjourney
Pic 7: Defendants' final design
Issues
In this case, the Zhangjiagang court focused on the following two issues: (1) whether Plaintiff's AI-generated pictures are copyrightable works protected under Chinese Copyright Law; and (2) whether Defendants constituted copyright infringement.
1.Whether Plaintiff's AI-generated pictures are copyrightable works protected under Chinese Copyright Law
The court held that Plaintiff’s AI-generated pictures are not copyrightable works and could not be protected under Chinese Copyright Law.
The court started its analysis that, if the content is primarily auto-generated by AI, then such AI-generated content shall not be copyrightable because it is the AI, rather than the user, who determines the output in the generative process. However, if on the other hand, the user takes AI merely as a tool and exercises his original intellectual input in the creation process, then the AI-assisted content should be protected as copyrightable work.
The court went on that, as to whether the user has exercised enough human creativity during the creation process, such records/documentation showing how the AI-assisted content were produced shall be provided. The court explained that, because it is difficult for users to create a particular expressive result through just a single round of prompts and parameter inputs, the original records that documented how the AI-generated content was created could prove whether users have made aesthetic choices and personal judgment during the generative process. Therefore, AIGC users shall provide clear documentation to prove that they have made personalized choices and substantive contributions to the expressive elements in the AI-generated content, by adding prompts and modifying parameters to adjust, select and refine the initial AI-generated output.
In this case, Plaintiff failed to provide enough documentation to prove her sufficient human creative inputs in the creation process. Meanwhile, Plaintiff admitted herself that it was impossible for her to regenerate the exact same AI-generated pictures, with the same prompts, due to randomness in Midjourney's generative process. Therefore, the court held that because Plaintiff failed to provide sufficient evidence to prove her original human authorship, the AI-generated pictures could not be protected as copyrightable works.
2.Whether Defendants constituted copyright infringement
Since Plaintiff's AI-generated pictures are not copyrightable and that Defendant's final chair products are not substantially similar with Plaintiff's AI-generated pictures, the court held that Defendants were not liable for copyright infringement.
In addition, the court held that the prompts used and shared by Plaintiff in this case, are merely ideas, and thus not copyrightable either.
In summary, the court held that Plaintiff's AI-generated pictures are not copyrightable and that Defendants did not constitute copyright infringement or unfair competition. This decision is final as ruled by the second-instance court.[2]
Comment
This is the first case where Chinese court denied copyright protection for AI-generated content for the past two years. What makes this decision interesting is that we have noticed similar rationales and findings between this decision and those demonstrated in similar U.S. cases.[3]
First, in the recent U.S. cases concerning the copyrightability of AIGC and the lately issued USCO report[4] on this issue, the U.S. Copyright Office and the U.S. courts have repeatedly focused on the issue "human creative control" in determining whether the AI-generated/assisted content is copyrightable. We have noticed similar discussion among Chinese courts on the issues of "human" and "creativity", but not so much on the issue of "control" (i.e., whether plaintiff/AIGC user could consistently, repeatedly and predictably produce the same AI-generated output), until this case. In this case, Defendants raised the argument that Plaintiff’s AI-generated content is "unpredictable", "uncontrollable" and "unrepeatable" and thus shall not be copyrightable. Plaintiff also admitted that she could not regenerate the same AI pictures as she posted earlier because of the "randomness and uncertainty" of the Midjourney AI service that she used. That is why the Chinese court concluded that because Plaintiff could not repeat her AI-creation process to create the same AI-generated pictures and that such creation process was not documented either, the court could not recognize Plaintiff’s personal selection and edits (original authorship) in the AIGC process.
Second, as with the USCO report (Part II) published earlier this year, the Chinese court also discussed the copyrightability of prompts, and concluded that, in general, simple text prompts are merely ideas and not expressions of ideas, and thus non-copyrightable.
Third, the Chinese court also required/encouraged AIGC users to "document" his AIGC creation process so as to prove sufficient human original authorship in the AIGC output. This approach seems to be similar to the "A Single Piece of American Cheese" case, where the USCO recognized the copyrightability of the AI-generated picture after examining the almost 10-minute video provided by the AIGC user/applicant.[5]
In summary, this case shows an interesting convergence with the recent U.S. cases concerning the copyrightability of AI-generated content. It would be interesting to see how this trend develops in other Chinese courts.
Thanks to Wang Mo and Zhao Yibing for their contributions to this article.
Scan the QR code to subscribe to "King & Wood Mallesons" for more information
See Feng v. Zhangjiagang Dongshan Cultural Commc’n Co., Ltd., the People's Court of Zhangjiagang City, Jiangsu Province, (2024) Su 0582 Min Chu No.9015, Mar. 19, 2025 (finding defendants not liable for copyright infringement and unfair competition).
See Feng v. Zhangjiagang Dongshan Cultural Commc’n Co., Ltd., Suzhou Intermediate People's Court of Jiangsu Province, (2025) Su 05 Min Zhong No.4840, April 17, 2025 (dismissing the appeal).
E.g. Letter from Robert J. Kasunic, Associate Register of Copyrights, Re: Zarya of the Dawn, to Van Lindberg, Taylor English Duma LLP (Feb. 21, 2023), available at https://www.copyright.gov/docs/zarya-of-the-dawn.pdf; and Letter from Suzanne V. Wilson, General Counsel and Associate Register of Copyrights, Re: Second Request for Reconsideration for Refusal to Register Théâtre D’opéra Spatial (Sept. 5, 2023), available at https://www.copyright.gov/rulings-filings/review-board/docs/Theatre-Dopera-Spatial.pdf.
See the United States Copyright Office, Copyright and Artificial Intelligence Part 2: Copyrightability (Jan. 29, 2025), available at https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-2-Copyrightability-Report.pdf.
See Invoke, Invoke Secures Copyright in Landmark Ruling for AI-Assisted Artwork (Feb. 10, 2025), available at https://www.invoke.com/post/invoke-receives-copyright-in-landmark-ruling-for-ai-assisted-artwork.
