News

The virtual digital person is a product of the digital era, which can be delivered naturally with real people, immersive experience, and interconnect the virtual world with the real world. There are also the following four major characteristics: first, digital attributes; second, identity attributes; third, visual presentation; and fourth, personified interactive attributes. The virtual digital person is an important asset for relevant enterprises to meet market demand, obtain trading opportunities and realise technological innovation. Therefore, after AI is becoming more and more popularised, there are more and more creators coming down and starting to use AI, to create some of their works.

This article will talk about one of the biggest concerns of many creators: AI infringement and copyright issues.

I. Series of Digital Human Infringement Cases

Virtual digital person kernel is divided into 2 kinds of drive: real person driven or algorithm driven, such as real person driven virtual man and intelligent driven virtual man. Virtual digital person belongs to the new things, and there is no law, judicial interpretation of the legal attributes of the virtual digital person for a clear characterisation. Therefore, the legal characterisation of virtual digital persons is difficult to determine. The biggest controversy is: whether virtual digital person has independent legal subject qualification?

The author will analyse the first infringement case involving ‘virtual digital person’ decided by Hangzhou Internet Court through cases to understand and analyse the issues related to the protection of intellectual property rights of virtual digital person in practice.

Case 1: The First Infringement Case Involving Virtual Digital Persons in China

In October 2019, a company released Ada, a virtual digital person, through a public event. the company comprehensively applied AI performance animation technology, ultra-realistic character intelligent modelling and binding technology, intelligent animation and voice synthesis technology, intelligent interaction technology and other artificial intelligence technologies to create Ada, an ultra-realistic virtual digital person. and in October and November of the same year, the company released two videos through a platform, one of which was used to introduce Ada, the other was used to introduce Ada. used to introduce the scene application of virtual digital person Ada, and one used to record the motion capture screen of real actor Xu and virtual digital person Ada.

In July 2022, a Hangzhou network company released two videos of the accused infringing videos through a Shake Yin account. The videos used relevant video content posted by a company in the centre of the video, and replaced the relevant logos in the beginning and end credits, and added marketing information about the virtual digital person course in the overall video. One of the videos also included the registered trademark of a Hangzhou network company and included the names of other virtual digital persons in the title of the video.

A company claimed that the above actions of a Hangzhou network company infringed on its right of information network dissemination of art works and audio-visual works, infringed on the right of information network dissemination of the video producer and the performer in the video production, and constituted an unfair competition of false propaganda, so it sued to the court, requesting the Hangzhou network company to eliminate the influence and compensate for the loss (including the cost of defending the right) of 500,000 yuan.

The Hangzhou Internet Court held that, of the two videos released by a Hangzhou network company, one video constituted an infringement of the right to network communication of audiovisual works, and the other video constituted an infringement of the right to network communication of works of art, video producers and performers, and that it had deleted the information content of the relevant logo in the video and replaced it with its own trade mark, which affected the rational decision-making of consumers and disturbed the order of market competition, directly affecting the rational decision-making of consumers. It also disrupted the order of market competition, directly harmed the commercial interests of a company, and constituted the unfair competition behaviour of false propaganda.

The three key points of this case are: firstly, whether Ada, a virtual digital person, can be the subject of copyright and neighbouring rights; and secondly, whether Ada can be the subject of copyright and neighbouring rights.

Secondly, whether from the perspective of enjoying rights and assuming responsibilities, or from the perspective of its existing degree of ‘autonomy’, the virtual digital person still could not become the subject of copyright; thirdly, whether the image of the virtual digital person Ada and the related video were the objects protected by copyright, and whether a company enjoyed the performer's rights of the video product in question.

The person behind the real person-driven virtual digital person is the essential participant. In this case, the ‘performance’ made by the virtual digital person is actually a digital projection and digital technology reproduction of the real person's performance, and it is not a performer in the sense of the Copyright Law, and does not enjoy the performer's right. When the virtual digital person participates in the filming or acts as a character, and his/her behaviour and performance activities are recorded and photographed on a certain medium to form a continuous moving picture, he/she also does not enjoy the copyright of audiovisual works or the neighbouring rights of video producers. Therefore, under the framework of the existing copyright legal system, the virtual digital person does not enjoy copyright and neighbouring rights.

In this case, the virtual digital person Ada is a real person-driven virtual digital person, not a real person modelling, i.e. it does not correspond to the digital doppelgänger of a particular natural person, and the process of generating the virtual image includes the creation of a static three-dimensional image, modelling and intelligent binding. The static three-dimensional image is then driven by a real actor (expression, movement, etc. capture) to show the dynamic visualisation effect, the virtual digital person Ada's facial expression and body movements can be presented in a way that fits the human body state. Xu meets the relevant provisions of the Copyright Law of the performer, as an employee of a company, is to perform, combined with the written agreement between the two parties, should be enjoyed by a company in the performer's right of property rights.

One of the videos released by a network company in Hangzhou constitutes an infringement of the right of information network dissemination of audiovisual works, and the other video constitutes an infringement of the right of information network dissemination of works of art, video producers and performers. The content of the video was not simply shared, but was used for the purpose of attracting traffic and marketing using the Shake video and Ada, the virtual digital person, which may affect consumers' rational decision-making, obtain more business opportunities and disturb the order of market competition, and therefore constituted unfair competition.

The case's denial of the subjective qualification of the virtual digital person's copyright and the affirmation of its own image as the object of copyright protection can, on the one hand, determine the rules for future disputes triggered by the relevant technology, and on the other hand, play a certain guiding role in socio-economic development and scientific and technological progress.

Case 2: The first AI-generated voice personality right infringement case in China

Another typical case is the first case of AI-generated voice personality right infringement in China on 23 April 2024, in which the plaintiff was awarded 250,000 yuan in damages in the Beijing Internet Court.

The plaintiff, Yin Mou, a dubbing artist, found that her voice was sold by an app called ‘Magic Voice Workshop’ under the name of ‘Magic Xiao Xuan’. Yin immediately sued the five defendants, including Beijing Intelligent Company, the operating body of ‘Magic Voice Workshop’, to the Beijing Internet Court on the grounds of infringement of the defendants' acts.

Defendant Certain Audiovisual gave the audio of three books recorded by the Plaintiff to Defendant Microsoft China, which then refined the Plaintiff's voice into an AI sound model and authorised the sale of it to Defendant Shanghai Certain Network and Defendant Beijing Certain Era. Then Beijing a time and small asked a sign a purchase contract, finally asked a buy Microsoft China's Yin AI sound model, shelves to the ‘magic sound workshop’. Although a certain audio-visual claimed that it had the copyright of the audio of the three books, it did not include the right to authorise others to use the plaintiff Yin's voice in the form of AI. Therefore, although it did not infringe on the ‘copyright’, it infringed on the plaintiff Yin's personality right.

According to Article 1023 of the Civil Code, the protection of the voice of a natural person refers to the protection of the right of portrait, and explicitly protects the voice as a special personality interest. That is, personality rights.

So the final judgement is, an audio-visual and Microsoft China has a clear infringement, to pay 250,000 yuan. The other three defendants belong to the sale relationship, because subjective fault does not exist, does not bear the responsibility for damages, so the judgement of the apology.

Case 3: AI Ultraman Infringement Case

Is it possible to infringe on content created directly by AI? AI is divided into service technology supporters (model developers), service providers (operators), and service users (users).

The most typical case is the case (2024) Guangdong 0192 Minchu 113 (AI Ultraman infringement case) decided by Guangzhou Internet Court.

The plaintiff was the copyright agent of Ultraman China, and the defendant was a company operating an AI website (alias Tab), which provided services with AI dialogue and AI-generated drawing functions. The Plaintiff alleged that the Defendant, through its website, generated images that substantially resembled images of Ultraman through AI. For example, when a user inputs ‘Generate an Ultraman’, an image of Ultraman is generated; and when a user inputs ‘Ultraman fused with American Girl Warrior’, an image of Ultraman's body spliced with an image of American Girl Warrior's long hair is generated. Moreover, the image of Ultraman generated by the website is substantially similar to the Plaintiff's image of Ultraman.

Intellectual property law has a particularly old principle is common sense, such as according to the common sense of individuals can judge the generated image, obviously is an infringing image, then constitute infringement. Unless it can be shown that no one at AI other than the Tab website knows about the Ultraman image and has never been exposed to it.

Although Tab is just a website that picks up other people's drawing APIs, according to Article 22(2) of the Interim Measures of the AIGC, ‘Generative AI service provider refers to organisations and individuals that provide generative AI services (including generative AI services through the provision of programmable interfaces, etc.) by utilising generative AI technology. ’ The Tab website was found to be a service provider (operator), constituting substantial infringement.

The final judgement was in favour of the plaintiff, and the court upheld the plaintiff's claims as well as in stopping the generation of Ultraman images. However, it did not support the plaintiff's request to delete the infringing data from the model training. The reason was that the defendant in the case was only the AIGC service provider rather than the service technology supporter (model developer), the large model was provided by a third-party company, and the plaintiff did not add the third-party company as a co-defendant.

However, according to Article 7 of China's AIGC Interim Measures, if you want to pass the algorithm filing, you must use data and underlying models with legal sources, and if intellectual property rights are involved, you must not infringe on the intellectual property rights enjoyed by others in accordance with the law. So for ordinary users whether infringement is actually clear, copyright is a very complete and unbreakable chain. Whether the material is in the model end, or in the middle of the data transmission, as long as there is a copyright problem, it will continue, will not be due to secondary processing or due to the base of a large number of users, so that the risk of copyright infringement has appeared to disappear.

Second, the digital human legal issues thinking extension

Artificial intelligence is a strategic technology that leads the future, is the core driving force of the new round of scientific and technological revolution and industrial change, and is considered to be the main position for the development of new quality productivity. The rapid development of artificial intelligence technology in China, the increasing abundance of data and arithmetic resources, and the continuous expansion of application scenarios have laid a solid foundation for the development of artificial intelligence scenario innovation.

Currently, a new generation of generative AI technology is being used by more and more people to create, generative AI technology makes people's way of creation change, which is the same as the impact of technological progress many times in history, the process of technological development is the process of gradually outsourcing human work to machines.

The Beijing Internet Court wrote in the judgement of the nation's first AI-generated voice personality right infringement case that ‘at this stage, generative AI models do not have free will and are not legal subjects. Therefore, when people use AI models to generate pictures, there is no question of determining who is the creator between the two subjects; essentially, it is still a person who uses the tools to create: that is, it is the person who makes the intellectual input in the entire creative process, not the AI model. Encouraging creativity is recognised as the core purpose of the copyright system. Only through the correct application of the copyright system, with proper legal means, encouraging more people to create with the latest tools, can we be more favourable to the creation of works and the development of AI technology. Against this background and technological reality, AI-generated images should be recognised as works and protected by the copyright law as long as they reflect the original intellectual input of a human being.’

The Guangzhou Internet Court said in its judgement in the AI Altman infringement case, ‘Considering that the generative AI industry is in the early stages of development, it is necessary to take into account both the protection of rights and the development of the industry, and it is not appropriate to overly aggravate the obligations of the service provider. In the process of rapid technological development, service providers should proactively and positively fulfil reasonable and affordable obligations of care, thereby contributing to the promotion of a Chinese-style AI governance system that is complementary to safety and development, balanced and inclusive, and compatible with innovation and protection.’

III. Conclusion

In conclusion, the author looks forward to a future of more perfect, comprehensive and flexible measures to promote technological progress, while protecting the legitimate rights and interests of creators, and better able to find a balance between service technology supporters (model developers), service providers (operators), and service users (users). All of us should be the beneficiaries of the development of new technology and new technology, as well as the practitioners to promote the healthy development of new technology and new technology under the legal framework.

Special Announcement: This article was originally written by the lawyers of JAVY law firm, it only represents the author's own views, and shall not be regarded as a formal legal opinion or advice issued by JAVY law firm or its lawyers. If you need to reproduce or quote any content of this article, please indicate the source.