Artificial Intelligence and Copyright in China: Lessons from a Recent Court Case

by Dennis Crouch

Thus far US copyright and patent tribunals have refused to award rights for AI generated works.  China has begun its move in the opposite direction with the recent decision granting rights to an artist who created an image using the popular generative AI system Stable Diffusion.  The case of Li v. Liu was decided by the recently Beijing Internet Court.  The court decides internet related cases using online tools. Although the court uses AI judges in some cases, this one was decided by humans.

[Read an English translation created by GWU Law Students Yuqian Wang and Jiaying Zhang supervised by Prof. Robert Brauneis: Li v Liu Beijing Internet Court 20231127 with English Translation. More AI related decisions are available at the GWU AI Litigation Database.]

The plaintiff, Li, used Stable Diffusion to generate a photorealistic image of an Asian woman. He shared the work on Xiaohongshu, a Chinese platform similar to Instagram. Later, the defendant Liu, a blogger, published the image in a blog post without Li’s permission, removing his username and platform watermark.

Li sued for copyright infringement. The Beijing Internet Court ultimately ruled in his favor, ordering Liu to apologize publicly, pay ¥500 ($72) in damages, and cover ¥50 ($7) in court fees.

Crucially, the court recognized Li as the legal rights-holder, having contributed the necessary “intellectual input” through choices of models, prompts, parameters, and final image selection. Judges emphasized that copyright law aims to encourage human creativity.

In this case, from the moment the Plaintiff conceived the disputed image to the moment the Plaintiff finally selected the disputed image, it can be seen that the Plaintiff has carried out a certain amount of intellectual input, such as designing the presentation of the character, selecting the prompts, arranging the order of the prompts, setting the relevant parameters, selecting which picture meets their expectations, and so on. . . . Currently, a new generation of generative AI technology is increasingly being used for creation. As long as AI-generated images reflect the original intellectual input of a person, they should be identified as a work and protected under copyright law.

This court ruling signals China’s interest in extending copyright law to better account for AI-generated content and also potentially places the country in the forefront of protecting AI creations.  Still, the Beijing Internet Court itself emphasized the approach of judging AI copyright issues based on the specific facts.

Of course, this is one regional lower court decision and is not generally a policy statement of the PRC.  The judges here particularly noted that the level of creative expression requires a case-by-case analysis.  In an email, Berkeley IP scholar Yuan Hao (郝元) points out, there is still debate within China on this issue. A similar case decided by a court in Shenzhen came to the opposite conclusion, finding AI-generated output eligible for copyright protection as the “intellectual fruits” of the AI developer rather than the AI user. According to Hao, cases like Li v. Liu feature “experimentalist practice” with different local courts exploring the issue before final guidance from the Supreme People’s Court.  Hao also disputes depictions of the case as a political move to strategically assist China’s AI industry. While acknowledging that some “political” IP cases may occur, she has not seen evidence of that here.  In my view, the low-value of the case (less than $100) also serves as an indication that this was not intended as a landmark decision setting national policy, but rather resolving a minor copyright dispute between two individuals.


41 thoughts on “Artificial Intelligence and Copyright in China: Lessons from a Recent Court Case

  1. 7

    I was wondering about the viability of the many new copyright suits against AI systems, including their database collections, but the downloaded copies of copyrighted materials alleged in the AI system output in the newly announce NYT suit do sound actionable. [Also, against anyone downloading that alleged AI query output, especially for commercial use]:
    This suit reportedly alleges that Microsoft and OpenAI copy Times content verbatim and even attributes false information to the Times. Also that prompts to ChatGPT asking it to reproduce paywalled content provided verbatim excerpts.

  2. 4

    The reasoning seems impeccable. We don’t award copryrights to paintbrushes either. An interesting issue is what happens when multiple “artists” use the same prompts on the same tools and get the same (close enough) results. Does the first one generated get the copyright?

    Should the models be trained to not produce any two images that are (x)(value) alike?

    Many people think “Chinese law” is an oxymoron because the upper levels of the government are lawless & so may dictate the outcome of any meaningful case?

    Does that make the entire body of law illegitimate? Reasonable people can disagree.

    The AI training question also does not seem to be a close case. If you buy a newspaper and feed it to your AI, what’s the damage to the copyright holder? What products weren’t bought or were bought from the wrong party?

    1. 4.1

      Meanwhile, the Thaler case continues to fuel the international debate about who to name as the inventor of a patentable invention created with help from an AI tool. I wonder, is Thaler bothering to run the issue also in China and, if so, what the outcome will be.

      1. 4.1.1

        … still waiting for your (Hobson) choice there on the change or not to the State of the Art.

        Beyond the simple binary answer, feel free to explain why you would choose that answer.

        If you dare.


          No use pressing me, anon, for an answer because I don’t understand your question. You will have to live in hope that somebody in your own jurisdiction i) understands you and ii) takes the time and trouble to formulate a reasoned reply.


            What exactly confuses you?

            It is a simple proposition: either the AI output advances the State of the Art or it does not.


              Do me a favour, anon. Give me the definition of “The state of the art” in your jurisdiction. I already gave you the definition under Article 54(2) of the European Patent Convention. If you don’t give me your definition, how shall I even begin to answer your question?

              There is a definition, is there not?


                Use the definition that you want to use, but DO answer the question, simply and directly.

                It is a simple – and direct – question.

                1. The definition I choose is that given by Art 54(2) of the EPC, namely:

                  “The state of the art shall be held to comprise everything made available to the public by means of a written or oral description, by use, or in any other way, before the date of filing of the European patent application.”

                  I would say that the output of an AI is a thing, and that the making available to the public of an output from an AI renders that output-thingy part of the state of the art relative to which patentability under the EPC falls to be assessed.

                  Where’s the catch? What is the trap that you think I have fallen into? Are you arguing that anonymous works don’t count as part of the state of the art?

                2. The “trap” as it were, is for those who seek to deny the capability of AI to be inventive.

                  The only way to move from a given State of the Art at time t1 to a NEW State of the Art at time t2 is for there to be an inventive addition TO the State of the Art.

                  Those (like Wt, but certainly not only him) who refuse to even discuss the matter and DENY that AI can be inventive, are stuck in the paradox that ANY addition by AI between times t1 and t2 CANNOT be inventive, and thus there can be no change in State of the Art, this — de facto — means that ALL output of AI is necessarily already a part of the State of the Art.

                  As I have posted previously:
                  AI aspects of invention are either:
                  1) BY AI, and thus no human can claim the element (an d any patents with AI claim elements thus is improper for no possible grant to an AI entity. This very much impacts (right now) co-inventorship status; OR
                  2) merely reflective of State of the Art, and thus is also not patentable. This very much impacts (right now) the entire obviousness inquiry.

                  There is no such thing as “AI is just a tool.” (mind you, truly generative AI)

                  Not in the patent sense.

                3. So now we have it. You assert this

                  “The only way to move from a given State of the Art at time t1 to a NEW State of the Art at time t2 is for there to be an inventive addition TO the State of the Art.”

                  The assertion is beyond my comprehension. You seem to be conflating novelty with obviousness. Who, other than you, does that? On what authority for it do you rely for your curious assertion?

                4. Not at all – as what I stated includes neither explicitly, but instead is the advance of the state of the art that encompasses just that: advances in the state of the art.

                  ALL else that does not advance the State of the Art must be within the State of the Art.

                  This is not difficult.


              Still not clear enough for me, anon. As to the “state of the art” I can understand what it means to “add” something to the state of the art but not what you mean by (variously) changing or advancing the state of the art.

              Perhaps your notion of what the expression “the state of the art” means is not compatible with what the EPC defines in Art 54(2) as “the state of the art”, namely everything that has been made available to the public? Perhaps you see “the state of the art” as some sort of interface between what is known and what is not. The peripheral envelope of an “expanding universe” as it were. That would fit with your fixation about “advancing” the “state of the art”. But that would mean that we are arguing past each other because we have different idas about what is the meaning of “the state of the art”.

              Germany used to have the idea that technological “advance” is a condition of patentability but it dropped that criterion ages ago as unworkable.

              Every day, the state of the art acquires more data, more testable propositions, more creative ideas, becomes denser. These things add to the state of the art and in so doing can render obvious that which was until then not obvious. Do they “advance” the state of the art though? You tell me.

              I persist with this discussion even though it appears pointless because it might help me in dialogue with the US patent attorneys who instruct me, if what they mean by “the state of the art” is whatever it is that you mean. So please do continue to enlighten me.


                You are trying WAY too hard – it is much, much, much more simpler than that.

                You tell me how adding to the state of the art is (somehow) not advancing the state of the art.

                Your attempts to NOT understand are driving you to lunacy.


                … as for your dealing with US counsel, you would be well advised that the full meaning of the word ‘promote’ in our Constitution’s delegation of authority to write patent law includes the nominal meaning related to “promotion,” as in, getting the word out.

                It is a mistake to think that ANY type of “advance (as in ‘improvement’) is a MUST. That most all cases may in fact be geared to that should ALSO be kept in mind (for those in the know that have studied innovation history and innovation management), that VERY OFTEN the breakthrough and industry-resetting advances are at first W0RSE than the status quo.

                These are NOT new things MaxDrei. I have provided them TO YOU on these boards many times now.

                1. You attribute to me a mind trying not to understand what you write. But when I seek clarification from you of what you wrote here:

                  “The only way to move from a given State of the Art at time t1 to a NEW State of the Art at time t2 is for there to be an inventive addition TO the State of the Art.”

                  observing that one can add to the state of the art without being inventive you reply that I’m making unnecessary difficulties and demand:

                  “You tell me how adding to the state of the art is (somehow) not advancing the state of the art.”

                  I see no value in continuing this dialogue, for you or for me. And there is by now nobody else following it.

                2. You attribute to me a mind trying not to understand what you write

                  There is no other viable reason when you CHOOSE to not understand the simple and direct words.

                  Your “proposition” that one can ADD to the State of the Art without being inventive is suspect – at best.

                  If one CAN add to the State of the Art without being inventive, then State of the Art has NO BUSINESS in patent law.

    2. 4.2

      Does the first one generated get the copyright?

      They all get a copyright. Copyright law does not work the same as patent law on this point. There is no novelty requirement in copyright law.


          The “author” is entitled to a copyright if he independently contrived a work completely identical with what went before; similarly, although he obtains a valid copyright, he has no right to prevent another from publishing a work identical with his, if not copied from his.

          Alfred Bell & Co. v. Catalda Fine Arts, 191 F.2d 99, 103 (2d. Cir. 1951)



            That Drum wants to accuse others of L Y 1 N G in ANY regards to the Hunter Biden fiasco is utterly beyond comprehension.

            If this is the Kool-Aid that Greg is ad d1c ted to, he will never recover.

    3. 4.3


      Without more (see 3.3 below), none of them get copyright.

      Any — thin — protection would only come for human creative items POST AI.

  3. 3

    Thus far US copyright and patent tribunals have refused to award rights for AI generated works.

    Is this correct? Can someone please cite me the example of the U.S. copyright refusing to register a work because the listed, human author(s) used an AI tool as part of the creative process? I can recall instances of refusal where the AI tool is listed as the author, but that is not really analogous to the Chinese case discussed above.

    1. 3.2


      Why hide Drum’s view of AI training? Is this not one of those links that actually may draw to the point of the article? (as opposed to the usual C R @ P that Greg uses ‘secret’ hyperlinks to signal whomever)…

    2. 3.4

      Yes, I do not see why this copyright registration by a human artist [not the AI] of an image generated using selective AI by a human artist is any different from U.S. copyright registrations of digital photographs or selective-computer-software-generated-artwork?

      1. 3.4.2

        There is no difference, Paul. So-called “AI” is a trendy buzzword for “computer” and only naive derps and propagandists like to pretend otherwise.


          LOL – “There is no difference, Paul. So-called “AI” is a trendy buzzword for “computer”

          That’s like the miscreant (and utterly 1gn0rant) old TV show “House” view that the first computer (somehow – maybe ‘magically’) contained ALL improvements thereof.

          But you be you Malcolm, and continue keeping your head up your arse.

  4. 2

    Your comment is awaiting moderation.
    December 28, 2023 at 7:44 am

    Lesson Number One:

    Chinese law is not US law.

  5. 1

    “… case as a political move to strategically assist China’s boring fake photos of people industry.”

    1. 1.1

      OT, Nice eye job.
      The SLANTS were happy with the way they looked. So why did their trademark cause such a problem. Maybe a real good trademark atty. could have shown they weren’t negating their look, but embracing it, by showing photos of other Asians that didn’t embrace their differences. It eould have been game over for the naysayers.

Comments are closed.