Recently, the Beijing Internet Court concluded a case that clarified the "work" attributes of images generated using artificial intelligence and the "creator" identity of the user. This is also my country's first copyright case involving images generated by artificial intelligence.
According to a notification from the Beijing Internet Court, the plaintiff used the open source software stable diffusion to generate the pictures involved in the case by inputting prompt words and then published them on the Xiaohongshu platform. The defendant published an article on Baijiahao, and the pictures involved in the article were used. The plaintiff believed that the defendant used the pictures involved in the case without permission and cut off the plaintiff's signature watermark on the Xiaohongshu platform, causing relevant users to mistakenly believe that the defendant was the author of the work, which seriously violated the plaintiff's right of signature and information network dissemination rights. , requiring the defendant to make a public apology and compensate for economic losses. The Beijing Internet Court made a first-instance judgment, ruling that the defendant apologized and compensated the plaintiff 500 yuan. Neither party filed an appeal. The first-instance judgment has now taken effect.
The judge responsible for hearing the case said that under the new technological background, traditional copyright theory no longer matches the reality of technological development, and should be adjusted and developed to adapt to changes in the actual situation and better meet the needs of rights protection and industry development needs.
01
Court: The picture involved reflects the plaintiff’s personalized expression
It is reported that the cause of this case was that the plaintiff Li used a large artificial intelligence model called stable diffusion to generate a picture by inputting prompt words. A picture of a girl in ancient costume, and published the picture on platforms such as Xiaohongshu under the name "Spring Breeze Brings Tenderness". The defendant Liu used the picture as an accompanying picture for the article on his Baidu Baijia account and removed the watermark on the picture. After Li found out, he sued Liu to the Beijing Internet Court for infringement of his rights of authorship and information network dissemination, demanding an apology from the defendant.
plaintiff Mr. Li initially claimed 5,000 yuan. He said that he put in a lot of labor when generating images. Although he only published a few works, he may have to operate and optimize hundreds of works in the background.
During the trial, both parties debated issues such as whether the pictures involved in the case constituted works, whether the plaintiff enjoyed the copyright to the pictures involved in the case, and whether the defendant's behavior constituted infringement. The plaintiff believes that he made a lot of intellectual investment in the process of generating pictures, such as designing the presentation of characters, selecting prompt words, arranging the order of prompt words, setting relevant parameters, etc., which reflects his originality and personalized expression. Therefore, the pictures involved are his own works and are protected by copyright law. The defendant argued that the main content he published was original poetry, not the pictures involved, and there was no intention to infringe. The pictures involved were obtained through Internet searches, and the specific source could not be provided, and it was not certain whether Li was using them. Have rights to the pictures involved.
After trial, the court held that the pictures involved in the case were graphic artistic works with aesthetic significance composed of lines and colors, and were considered works of art under copyright law. The pictures involved in the case are original and reflect people's original intellectual investment. They should be recognized as works and protected by copyright law. The plaintiff is the one who directly set up the artificial intelligence model involved in the case as needed and finally selected the pictures involved in the case. The pictures involved in the case were directly generated based on the plaintiff’s intellectual investment and reflected the plaintiff’s personalized expression. Therefore, the plaintiff is the owner of the pictures involved in the case. The author owns the copyright of the pictures involved. The defendant's behavior in this case infringed upon the plaintiff's right of authorship and information network dissemination rights for the pictures involved in the case, and he should bear infringement liability. Taking into account the specific circumstances of the case and the circumstances of the infringement, the court finally ruled that the defendant should apologize to the plaintiff and pay 500 yuan in compensation for economic losses.
Beijing Internet Court judge Zhu Ge, who heard the case, said that during the trial of the case, they repeatedly considered whether traditional theories should be adapted and developed when encountering new application scenarios, and whether the identification of works is only legal. Judgment, or value judgment is also required. Zhu Ge pointed out that only by adhering to a future-oriented judicial concept can we better encourage the application of new technologies and promote the development of new business formats.Since entering the era of artificial intelligence, human creation tools have undergone fundamental changes. People no longer need to draw lines and fill in colors with their hands, but use AI to create. However, this does not mean that humans do not need to do some processing of picture elements. Select and arrange. In this new technological context, traditional copyright theory no longer matches the reality of technological development, and should be adjusted and developed to adapt to changes in the reality and better meet the needs of rights protection and industrial development.
In addition, Zhu Ge believes that countries around the world regard "originality" as the core component of defining works, but they have not clearly defined or explained it in legislation. By recognizing the "work" attribute of artificial intelligence-generated images and the user's "creator" identity, it will be helpful to encourage users to use AI tools to create, thus achieving the inherent goal of "encouraging the creation of works" in copyright law and promoting the promotion of Relevant entities use AI-generated content to label content to promote the implementation of regulatory regulations and the protection of the public's right to know, which is conducive to strengthening the dominant position of humans in the development of artificial intelligence and promoting the innovative development and application of artificial intelligence technology. Should copyright protection also apply to
02
training data? What does
think about this case? Wang Congwei, a lawyer at Beijing Jingshi Law Firm, told Nandu reporters that as works, AI-generated images are qualitatively no different from works without AI participation, because when determining what kind of works constitute such images, the court believes that they do not qualify as "qualified works". other intellectual achievements characteristic of the work”, but are works of fine art. However, since AI-generated images are different in originality and use methods from works without AI participation, there may be some differences in the rights of authors.
According to Article 54 of the Copyright Law: “In case of infringement of copyright or copyright-related rights, the infringer shall compensate the right holder in accordance with the actual losses suffered by the right holder or the infringer’s illegal gains; If the illegal gains of the right holder are difficult to calculate, compensation may be given by reference to the rights usage fees. If the actual losses of the right holder, the illegal gains of the infringer, and the rights usage fees are difficult to calculate, the People's Court shall, based on the circumstances of the infringement, make a judgment to award a compensation of more than 500 yuan. Compensation of less than five million yuan."
In this case, the court finally determined that the amount of economic loss that the defendant should compensate the plaintiff for the alleged infringement was 500 yuan based on the circumstances of the pictures involved and the circumstances of the infringing use. Wang Congwei said that the amount of compensation awarded by the court in other infringement cases before this case has certain differences. The amount of compensation is generally determined based on the actual situation of the case. The comprehensive consideration factors include the benefits obtained by the infringing party, the creation of the infringed party The cost, the value of the work itself and the social value of the work itself, etc. For example, for picture infringement cases that are relatively large in terms of number of cases, the general amount of compensation is 800 yuan per picture. "Compared to this case, the basis and amount of compensation are not significantly different from other types of cases, which shows that such cases only include the AI factor in the basis of the creation of the work, but the legislative purpose of the copyright law is still maintained and adhered to when handling it. , is to protect the originality of the work."
Wang Congwei told Nandu reporters that this case is of great significance for the protection of intellectual property rights in works created using artificial intelligence technology in the Internet era. She pointed out that the qualitative issue of artificial intelligence-generated products in the field of intellectual property has always been the focus of academic debate, and this case affirmed the originality of AI-generated images and determined that images generated by generative artificial intelligence have the attributes of works and can be obtained Copyright, and recognizes that artificial intelligence users enjoy copyright as authors in accordance with the law. This precedent will have a certain impact on subsequent judicial practice, provide a reference precedent for similar cases, and provide certain guidance for the handling of similar disputes in the future.
Xu, director of the Digital Economy and Legal Innovation Research Center at the University of International Business and Economics, holds a similar view.He said that our country is not a country of case law, so the judgment of the Beijing Internet Court cannot be directly used in subsequent cases. However, given that there are currently only three Internet courts in the country, the cases of the Beijing Internet Court have played an important leading role in many cases. This This may lead to the case being widely cited.
In addition, Licensing also pointed out that this case may have a profound impact on the industry. Since there is a copyright issue with training data in aigc products, if a work produced using aigc is judged to be copyrighted, should the training data also be subject to copyright protection requirements? Licensing believes that interests can only be balanced by ensuring that the training data for AIGC products is authorized by the copyright owner.
He believes that in this case, in addition to paying attention to the basis of the judgment, we should also consider its subsequent impact. “It may bring about the issue of rebalancing interests in related industries. For example, more rights holders may be involved in the future. File relevant rights claims on the input side of AIGC (i.e., training data), because since the court has determined that the output of AIGC can be protected by copyright, of course the input side must also claim copyright protection. Therefore, from a compliance perspective, among them There is a problem, that is, the platform is legally obliged to prove that it does not infringe the intellectual property rights of others."
Written by: Nandu reporter Hu Gengshuo
