Artificial Intelligence generatives can be used as work to protect existing judicial cases. The characteristic of artificial intelligence in copyright law is that it is less or no direct intervention by humans, and products that are less directly intervened by humans can be protected by copyright law in precedents.
On the premise that artificial intelligence generators are recognized by works, they give artificial intelligence creators the copyright of artificial intelligence generators, and then construct a platform or outsourcing artificial intelligence generator rights utilization model through rights transfer and rights adjustment.
Author | UCID Technology Co., Ltd. Sun Wen
Edit | Youqing
1. The main reason for the conflict judgment is that the judgment is analyzed
0 Copyright infringement case of Yingxun (hereinafter referred to as the "Tencent Case") was concluded in the first instance in the Nanshan District People's Court of Shenzhen. This case is the first time that artificial intelligence genes constitute works in the sense of copyright law. This is also the first time that judicial practice has responded to whether artificial intelligence genes can constitute works.
However, in the verdict of Feilin Law Firm v. Baidu (hereinafter referred to as the "Filing Case") in May 2020, the Beijing Intellectual Property Court determined that the graphics automatically generated by the customized graphics tool do not constitute a work in the sense of copyright law. The difference in the previous and subsequent judgments caused controversy.
Whether artificial intelligence products can form works have been discussed in the academic community, and the different understandings in judicial practice have given new inspiration to research.
Comparing the two judgments, it is not difficult to find that the court has different understandings and intentions in different utilization modes of artificial intelligence and different participants generated by artificial intelligence products.
In Tencent case, Tencent is both the creator of artificial intelligence and the user of artificial intelligence. Its artificial intelligence is directly used by itself after creation, and the behavioral pattern reflected is the process of individuals creating artificial intelligence as props [1].
In the Film case, Film Law Firm is a user of artificial intelligence, and its behavioral model is to use the implementation of settings on the artificial intelligence creation platform provided by third parties to obtain the content they need [2]. The utilization model of these two AIs also reflects the industry transformation model provided by AI from its own use to the platform [3].
combined with the two judgments, the conclusion can be drawn that assumes that artificial intelligence products can constitute a work, so the court is willing to give AI creators rather than other participants a certain exclusive right in a copyright model [4].
Based on the two cases, this article will try to include the results of the judgment in the existing discussion, but this process first needs to solve the problem of how to accept the judgment in the Tencent case, and secondly, it is to solve the problem of how to construct a rights allocation model that conforms to the industry model on the premise that the product constitutes a work. This article will be developed with this idea.
2. Correctly understand the elements of "no man-made intervention"
Artificial intelligence itself is a product of technology, expressed as the research and development of theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. The definition of technology or discipline is often a refinement of the difference between itself and other technologies, and cannot reveal the characteristics of artificial intelligence in law, especially copyright law.
Scholars often classify artificial intelligence into strong artificial intelligence, weak artificial intelligence and super artificial intelligence based on the different development levels of artificial intelligence, and then judge whether it has the "original" ability to determine whether its products constitute a work. Classification such as
has certain technical descriptive significance, but it is not appropriate to determine the legal characterization based on technical standards.
On the one hand, the division of strong and weak artificial intelligence in terms of technology has no natural reference value in copyright law [5]. On the other hand, it is of course certain that the so-called strong artificial intelligence has the subject qualification, which makes some artificial intelligence products that do not have the necessary protection have unnecessary rights.
Copyright Law for artificial intelligence should try to pay attention to new content in addition to having the "original" ideas that may produce similar human wisdom.
wipo dissolves the concept of artificial intelligence into a machine or system that can perform tasks that are considered to be completed by human intelligence without finite or no human intervention [6]. From the above definition, it is not difficult to find that Wipo believes that the fundamental feature of the new copyright law of artificial intelligence is "less or no human intervention". The idea of
is related to the characteristics of its "original" advocated by scholars. Describing artificial intelligence without intervention is based on the aspects of human beings as the subject and machines as the object, while the "original" of artificial intelligence is based on the positive aspects of machine as the subject. Emphasizing that artificial intelligence can complete creation alone is two of a behavior In terms of description, the two have a relationship of one and two sides.
has little or no human intervention. As artificial intelligence, it is necessary to accurately understand the meaning of "intervention" in copyright law. It should be believed that although artificial intelligence itself has gradually deviated from the category of "tools", the artificial intelligence encountered in judicial disputes or in real life are designed and created to achieve human goals.
Although artificial intelligence generators are not directly created by humans, the purpose, path and even raw materials for producing the generators are from humans themselves.
In this sense, artificial intelligence is similar to the evolution of human tools in copyright law.
Human The creations are all based on tools, from relatively primitive pens to later cameras and cameras. When copyright law requires that works be fixed on a certain medium, human creation cannot be separated from tools. Traditional works are extensions of tools by people, that is, the direct person who generates works is a tool used by people, but they ultimately come from human thinking and decisions.
In the process of emphasizing that artificial intelligence is less or not interfered with, it means that artificial intelligence products are automatically generated by software designed by software designers. Artificial intelligence products have no contact and intervention when they are automatically operated by artificial intelligence.
But the established rules, algorithms, and designs of artificial intelligence itself reflect the designer's intervention and choice, that is, the design of the software or system itself reflects the designer's "originality" Value.
emphasizes the non-intervention nature of artificial intelligence products, which means that there is no human intervention in the process of artificial intelligence products, rather than absolutely no human factors. In the context of the inability to evolve or the arguments of artificial intelligence cannot evolve or conflict with each other, based on the characteristics of artificial intelligence without human intervention, whether the product can have certain new value as a work.
3. Case analysis of works produced without human intervention
As a product without human intervention, the representative and origin of the case of the work that constitutes a work. The monkey photo case in the United States is a powerful refutation of the artificial intelligence products.
The US court believes that the photos taken by monkeys by chance cannot constitute a work because the subject taken is not a human. If the photo is determined to constitute a work, it will undermine the consistent tradition of the copyright subject being a human being required by the US Copyright Law .
As an animal, monkeys cannot form the author according to the idea of "human-creation-work", and their "creation" photos do not constitute works under copyright law. Like monkeys, artificial intelligence is not a human subject and cannot become an author based on the idea of creating works, but this does not mean that artificial intelligence products cannot form works like monkey photos.
The subject of the monkey photo shoot is a monkey, which is not interfered with or controlled by the human subject. The photos involved are neither created by people themselves nor created by human control or influence. From this perspective, the accidental photography of monkeys cannot constitute a work.
The artificial intelligence currently involved is a machine or system generated by humans. From the perspective of the controversy, the current artificial intelligence products are automatically generated by the artificial intelligence system after operating or simply providing materials for a certain purpose or design. Since humans have a certain degree of intervention in artificial intelligence itself, whether it is using artificial intelligence as an auxiliary tool or thinking that artificial intelligence generates works by itself, the products have more or less connections with the designer. This is the difference between animal-related products and artificial intelligence products [7].
In addition to "works" related to animals, automatic video screenshots are also considered to be a typical case where there is less human intervention. Medical membrane photo cases are earlier cases of automatic videos and screenshots.
In this case, the court held that although the screenshot of the plaintiff was taken from the operation video, the plaintiff took the key images of the clinical application of the medical membrane from the surgical video he performed and reflected the originality of the intellectual labor that reached the minimum required by copyright law, so the screenshot constituted a work. [8]
In this case, the surgical video recording is a single stylized result of the whole process, without any artificial control, or even with artificial control and intervention, this choice is purely a choice from angles made for the convenience of shooting, and this choice cannot reach the height of originality.
But from the judgment results, can meet the criteria of originality even in videos that do not reflect originality, which initially shows that the products taken by mechanically have the possibility of constituting the work.
This tendency is more obvious in Gaoyang’s case of infringement of Youku’s high-altitude photography works.
video screenshots in the case
In this case, the court further determined that the automatically captured video is original, and then determined that the screenshots of the video constitute the photographic work should be protected by the copyright law. Specifically, the court held that "for those shootings that reflect manual intervention, selection and have clear purpose, even if they are mainly completed automatically by machines, as long as they meet a certain artistic quality, it cannot be denied that they can constitute a work."
It should be pointed out that the "artificial intervention" mentioned by the court here refers to the "human factor" mentioned above, because during the entire video recording process, people are not involved in the video recording process at all, but only human factors such as the position of people before, which continue to play a role. Based on this idea, the court extracted the purpose of the shooting, subject, shooting techniques, shooting equipment, etc. that can reflect the human factors involved in the case being a shooting process, and finally came to the conclusion that the video of the case constituted the work. [9]
Computer continuous pictures are the "audio-visual works" closest to the expression form of artificial intelligence generics. In the case of NetEase v. Huaduo copyright infringement, the court determined that the continuous picture of the game constituted an electronic work.
The court held that the game involved, as a computer software program, was a comprehensive audio-visual expression of the game automatically or in response to the interactive instructions of the game user. The overall picture of the game reflects the game developers' comprehensive considerations of story system, gameplay rules and overall artistic style, as well as the organic combination of game materials, which obviously reflects the game developers' personal choices and arrangements. [10]
Similarly, in the case of Feihu v. Baiji copyright infringement, the court pointed out that "in automatic hanging-up game, the game engine must follow certain game rules when calling game materials to form a continuous picture of an organic whole." [11] This also shows that the court believes that even if the final picture generated by the computer is automatically generated under the guidance of the rules, the picture still constitutes a work.
It is not difficult to see from the above cases that the autonomy of human creation has gone from low to high, but the degree of intervention of humans in tools or carriers is getting higher and higher.The videos automatically captured by
can only complete a single recording function, while the computer can realize the operation of multiple programs to generate audio-visual effects. Correspondingly, automatic shooting can only intervene in the camera with a certain degree of intervention in parameters such as angle and duration. The continuous machine images generated by the computer are audio-visual results obtained by the programmer and in all aspects, and people have higher control and choices of computer software or tools.
Compared with artificial intelligence, when completing an artificial intelligence, people not only need to call various resources, but also need to design rules, algorithms, and templates to enable artificial intelligence to obtain certain computing and learning capabilities.
The emergence of artificial intelligence shows a trend consistent with the previous tool development model, namely, the technical idea of increasing human intervention in tools and increasing autonomy of machines. Combined with the judgment results of Tencent v. Huaxun, the new judgment determines that artificial intelligence products constitute works in the sense of copyright law, in line with existing judgment ideas.
4. The artificial intelligence designer is the author of the product
When determining that the artificial intelligence generator constitutes a work, it is necessary to further clarify the author of the artificial intelligence generator. In the Tencent case, the court believed that Tencent, as the designer and user of artificial intelligence, certainly enjoys the copyright of the product. The court did not refuse or directly respond to whether the machine could constitute the author of the work, but instead identified the designer as the author by the generated objects created by the legal person, reflecting the designer's thoughts and choices, thereby avoiding the aforementioned questions. Answers like
reflect the existing consensus among scholars. Although the academic community has different choices for the protection path of artificial intelligence, they often believe that the products should be protected by copyright law, and at the same time, they also refuse that artificial intelligence can become the subject of copyright law [12].
When the designer and user of the artificial intelligence platform in the Film Case were separated, the court refused to determine that the automatically generated icon was a work, and also believed that the product should be protected by using the anti-unfair competition law path [13]. This has reached a consensus that rejecting artificial intelligence as the creative subject and need to give certain incentives to artificial intelligence producers.
examines extraterritorial law enforcement and judicial practice. Rejecting artificial intelligence as the creative subject is also the mainstream idea of European and American countries [14], but refusing artificial intelligence as the creative subject does not mean refusing to grant certain rights to artificial intelligence designers, that is, the interests of copyright and even patent rights can be enjoyed by artificial intelligence designers [15]. Therefore, after giving up the idea of creating works by machines [16], it is more appropriate to identify the designer of artificial intelligence as the author of the product.
5. Protection path construction of artificial intelligence generators
If the artificial intelligence generator is given the status of the work, the artificial intelligence designer can, of course, protect their interests by exercising the exclusive rights granted by the copyright law. However, when the artificial intelligence designer is separated from the user, it is necessary to ensure that the user has sufficient rights to control the work.
As the number of artificial intelligence users is increasing and even becoming more and more an artificial intelligence platform has become a model provided by artificial intelligence, artificial intelligence users often have more direct interest demands for artificial intelligence products. According to the basic idea of how much it contributes to the product, it is the most basic rule to assign initial copyright to the artificial intelligence designer, but if the rights are completely attributed to the artificial intelligence designer, it will hinder the use of artificial intelligence users and their control of the improper use of the product by third parties.
At this time, the designer's copyright of the product to the user of the product can be transferred to the user of the product through legislative rights transfer, and at the same time, certain adjustments are made to the copyright ownership rules [17], so that the user can enjoy certain exclusive rights on the product.
In this way, artificial intelligence designers can obtain incentives for certain products through license distribution and sale to increase costs or revenue commissions; while artificial intelligence users can produce different works by using artificial intelligence to obtain benefits through the exclusive rights given by copyright law, thereby promoting a positive cycle of generation and utilization in the industry.
It should be noted that when determining that the product constitutes a work, the product should first satisfy the "original" appearance. Although there is a certain difference between the expression of artificial intelligence products and existing works, it should be believed that if all differentiated artificial intelligence products constitute the work, some content that should not constitute the work will become obstacles to gaining profits and undermining the prosperity of the copyright market.
Therefore, in judicial practice, when examining whether artificial intelligence products can form works, the court should conduct an investigation in the form of case analysis and strict originality standards to prevent unnecessary artificial intelligence products from being added to the ranks of the works.
Conclusion
Artificial intelligence has indeed entered people's lives, and the judicial cases that actually occur clearly show this. Compared to the question of whether the era of strong artificial intelligence will eventually come, the issues that need to be solved before the judicial department and the establishment of judgment rules are the top priority.
Determining that artificial intelligence products can constitute the evolution of the work in accordance with the case of no human intervention, but at the same time, it should be recognized that defining artificial intelligence products as works can promote the development of the artificial intelligence industry on the one hand, and on the other hand, it may also cause the phenomenon of crowding out public resources due to a large amount of artificial intelligence creation, and even causing difficulties in creating materials for natural persons. How to further balance the contradiction between the two requires further waiting for the industry to give new answers and inspiration.
