On May 29, 2024, a research team at Stanford University released a model called "Llama3V", and someone disclosed on Twitter that the Stanford AI team plagiarized the Chinese open-source model. Someone asked in the media: "Is it infringing for the Stanford AI team to plagiarize China's wall-facing open source model?" The reason for this is that there are similar developments in the mainland that have been "stillborn" due to the "exposure" of a certain self-media, and whether plagiarizing open source software is infringing needs to start with specific cases.
Today's Legal Q&A: Copyright
1. Mulan's "skinning" Python incident
In January 2020, a domestic self-media disclosed: "The Chinese Academy of Sciences made a fuss, and Python 'changed its skin' into Mulan." The cause of the incident is: After the release of the article "Facing the Intelligent Internet of Things, the Chinese Research Team Released the "Mulan" Programming Language System", someone opened the Mulan language environment and found that many functions of the command-line programming environment were consistent with Python.
After a few days of fermentation of the Mulan "skinning" Python incident, on January 19, Liu Lei, the head of the "Mulan" language research and development team, said "solemn apology", and the self-media that disclosed the incident gained a lot of traffic. Subsequently, the Chinese Academy of Sciences responded to the "Mulan" plagiarism Python incident and suspended the party Liu Lei for inspection. Some people may ask, why did the copyright owner of the original Python not raise an objection when the Chinese scientific research team released the "Mulan" programming language system, but the Chinese were confused?
The reason why the Chinese are in a mess is simple. On the one hand, the Copyright Law stipulates criminal liability, while judicial practice does not understand the provisions of Article 53, that is, the pursuit of criminal liability is premised on administrative violations, specifically "infringement and damage to the public interest". On the other hand, most people understand that computer software cannot systematically explain the protection of such rights, for example, Article 64 of this Law stipulates that "the measures for the protection of computer software shall be separately formulated by the State Council." Article 3(8) includes computer software, and many people do not see the aforesaid provisions.
Mulan "skinning" Python event
2. The consequences of China's lawsuit against the Stanford AI team
According to the theory of civil procedure, or rules, there are two ways for China to sue the Stanford AI team for infringement: one is to choose to sue in the United States. The second is to choose to sue in Chinese courts. According to the current judicial practice, or judicial interpretation, when I sue the Stanford AI team for infringement, the mainland court will generally not accept it. It may be argued that the inadmissibility of domestic courts is incompatible with the provisions of the Code of Civil Procedure.
Article 530 of the Civil Procedure Interpretation clearly stipulates that if a foreign-related civil case meets the following circumstances at the same time, the people's court may rule to reject the plaintiff's lawsuit and inform the plaintiff to file a lawsuit with a more convenient foreign court: (1) the defendant submits a request that the case should be subject to the jurisdiction of a more convenient foreign court, or raises an objection to jurisdiction; …… (6) The foreign court has jurisdiction over the case and it is more convenient to hear the case.
Some people may ask, what is the result of China's lawsuit against the Stanford AI team for infringement in the United States? The United States is an unwritten law country that determines whether computer software is infringing by precedent. English judges have long been jurisprudence for computer software infringement: that is, the protection of computer software does not involve ideas and means of expression. Accordingly, most of the results of China's lawsuit against the Stanford AI team in the United States did not constitute infringement.
3. Reasons for the other provisions of the Copyright Law on computer software
Article 3 of the Copyright Law stipulates eight specific types of copyright rights, and computer software rights are placed in the last item, i.e., item 8. The reason for this is that many types of civil law countries do not protect computer software as copyright, for example, in continental European countries, the aforementioned precedent is that the United Kingdom does not have a pre-Brexit judicial decision.
The expression of computer language is limited, only "0" and "1", according to which, the self-media person who found Mulan's "skinning" Python incident is really "talented": judging from the many "0" and "1" "skins". "0" and "1" express linguistic information is not an invention of computer people, for example, the Morse code in the telegraph is a combination of "short" and "long".
If "0" is considered to be "short" and "1" is "long", can the inventor of Morse code sue the computer software program writer for infringement? Of course not, the reason for this is that even if Morse code is protected by copyright law, its term of protection has long since expired.
Article 64 of the Copyright Law authorizes the State Council to provide for the protection of the right to disseminate computer software and information on the Internet. In 2001, the State Council promulgated the Regulations on the Protection of Computer Software. Article 6 stipulates that "the protection of software copyright in these Regulations shall not extend to the ideas, processing processes, operating methods or mathematical concepts used in the development of software". The "etc" make software infringement lawsuits extremely rare in China.
Article 29 further stipulates that: "Software developed by a software developer that is similar to pre-existing software due to the limited number of expressions available does not constitute an infringement of the copyright of pre-existing software." "This article is, in fact, a statement of the rationale for the limited protection of computer software.
As far as the Stanford AI team's plagiarism of China's wall-facing open source model is concerned, its particularity is also "open source". Open source should be understood as the public information or basic information of computer software, for example, many software is developed based on open source information to achieve the purpose of "compatibility", or requirements. Accordingly, in terms of computer software development, the relevant teams in the mainland need to have the courage to "take it", otherwise, Chinese software developers may watch the progress of the software systems of the Americans and Indians, for example, there is no popular computer operating system in the mainland.