In a recent case decided by the Beijing Internet Court, the plaintiff claimed that the defendant had infringed his copyright by using an image he had created using the Stable Diffusion generative AI model without his permission. The plaintiff had used the model to generate the image by inputting prompt words, and had posted it online with a watermark. The defendant had allegedly removed the watermark and used the image on her own social media account.
The court considered several key factors in reaching its judgment. First, it examined whether the image created using the Stable Diffusion model constituted a work of art and was eligible for copyright protection. The court found that the image met the definition of a work and was eligible for copyright protection, as it reflected the plaintiff’s original “intellectual investment” in selecting and using the model to create the image.
In particular, the court affirmed that the “image is not a call to an existing ready-made image through a search engine, nor is it an arrangement and combination of various elements preset by the software designer. In Iayman’s words, the role or function of this model is similar to how humans have acquired skills and abilities through learning and accumulation. lt can generate corresponding images based on text descriptions entered by humans, draw lines and colors on behalf of humans, and integrate human activity, ideas are presented in a tangible way. In this case, the plaintiff wanted to draw a close-up of a beautiful woman in a photographic style under dusk light conditions, and then entered the prompt words into the Stable Diffusion model.”.
Second, the court considered the issue of copyright ownership of the image, finding that although the AI model had generated the image, it could not be considered the author or owner of the copyright. The designer of the AI model had only produced the creative tools, and the intellectual investment was reflected in the design of the model, not in the specific image generated. Therefore, the plaintiff was the author and held the copyright of the image.
About this, the court further affirmed that “the plaintiff was the one who directly set up the artificial intelligence model involved in the case as needed and finally selected the pictures involved. These were directly generated based on the plaintiff’s intellectual investment and reflected the plaintiff’s personalized expression.”
Third, the court examined whether the defendant had infringed the plaintiff’s copyright by using the image without his permission and removing the watermark. The court found that the defendant had indeed infringed the plaintiff’s copyright and ordered her to issue a public apology and compensate the latter for economic losses.
It is interesting to note how the Beijing Court makes a few considerations about the technology at stake, in order to reach to its final judgment, which are worth reading as they are written in the decision.
“The Stable Diffusion model and models with similar features can generate beautiful images based on text descriptions. Many people, including those without drawing skills, are trying to use these new models to generate content and present their creativity and designs in a tangible way, greatly improving the efficiency of image creation.
It should be said that general artificial intelligence technology has changed the way people create, which is the same as the impact of many technological advances in history. The evolution of technology is the process of gradually outsourcing human work to machines.
Before the advent of cameras, people had to use superb painting skills to reproduce the image of objective objects. The advent of cameras made it easier to capture the image of objective objects. Now, the camera functions of smartphones are becoming more powerful and easier to use. However, as long as the photos taken with a smartphone reflect the photographer’s original intellectual effort, they still constitute a photographic work and are protected by copyright.
It can be seen that the more technology develops, the smarter the tools, the less people invest, but this does not affect our continued use of the copyright system to encourage the creation of works.
Before the advent of the above model of artificial intelligence, people had to spend time and energy to produce certain drawings. In order to obtain a painting, it is necessary to have skills or to hire others. In the scenario of entrusting others to paint, the client will provide certain requirements, and the trustee will draw lines and fill in colors according to the client’s needs to complete a work of art. Between the client and the trustee, generally speaking, the trustee who puts pen to paper is generally considered the creator.
This situation is similar to the situation where people uses artificial intelligence models to generate images, but there is one major difference between the two, that is, the trustee has his own will. When he completes the painting commissioned by the client, he will own ideas into the painting. Trade-offs and judgments.
At this point, the generative AI model lacks free will and the legal subject status. Hence, there is no issue in deciding the creator between two parties when individuals use AI models to generate images. They are simply using tools, with people being the intellectually involved party throughout the entire creation process, not the AI model author.
Encouraging creation is recognized as the core purpose of the copyright system. By correctly applying the copyright system, using appropriate legal means, and encouraging more people to use the latest tools to create, we can better promote the creation of works and the development of artificial intelligence.
Under this background and technological reality, as long as images generated by artificial intelligence can reflect people’s original intellectual investment, they should be recognized as works and protected by copyright law.”
Header image generated with Microsoft Bing; prompt: “It should be said that general artificial intelligence technology has changed the way people create, which is the same as the impact of many technological advances in history. The evolution of technology is the process of gradually outsourcing human work to machines.”
