Life

Who does the copyright of AI-generated works belong to?

  It is fun to use AI software to generate various articles, images, and videos, but when it comes to copyright, it is unclear for a while. Whether these “works” are original works, and who owns the copyright, these issues need to be explored urgently.
  Since the end of last year, the chat generation pre-training transformation model ChatGPT has gradually become popular on social networks, and content generation AI software such as Midjourney and Dall-E have also attracted the attention of many netizens.
  Are things that people “create” with AI tools protected by copyright? Who does the copyright belong to? There is no consensus in the legal community.
  With the rapid development of AI technology in China in recent years, there have been some classic cases that have focused on different aspects of the copyright issues of AI works.
  Case 1: Without originality, there is no
  copyright ——Movie Volume · Beijing Chapter”, the text contains diagrams. The day after it was published, a very similar article appeared on Baidu Baijia. The first and last paragraphs were removed, but the diagrams in the article were retained. Therefore, Film Law Firm sued Baidu Company, arguing that Baidu Company published the article involved in the case on the Baijiahao platform without permission, infringing its copyright, and requested that Baidu Company be ordered to make an apology, eliminate the impact and compensate for economic losses.
  The defendant, Baidu, argued that the article involved in the case of Film Law Firm was not original in itself, and was not created through its own labor, but was generated by using legal statistical data analysis software, so it did not fall within the scope of protection of the Copyright Law. As long as the film law firm searches keywords in a certain database and uses the “visualization” function, it can automatically generate an analysis report including charts and text. Baidu believes that Film Law Firm is not a qualified subject in this case, and there is no evidence to prove that the article involved in the case is the work of a legal person.
  After two trials, the Beijing Intellectual Property Court finally did not support the rights protection request of Film Law Firm. The court held that the graphic part of the article involved in the case was produced using relevant software. Although it may appear in different shapes due to data changes, it is only based on data differences rather than creation, so it cannot reflect originality. Film Law Firm claimed to artificially beautify the lines and colors of the graphics, but did not submit evidence to prove it. Therefore, the court held that the graphics in the article involved did not constitute graphic works. Since this software product is not a work, there is no copyright protection.
  According to Article 2 of the Implementing Regulations of my country’s Copyright Law, the term “works” in the Copyright Law refers to intellectual achievements that are original in the fields of literature, art and science and can be reproduced in some tangible form. A work under copyright law must meet four elements: it belongs to the field of literature, art and science; it is original; it can be reproduced in some tangible form; it is an intellectual achievement.
  It can be seen that works under the copyright law cannot be separated from the creative factors of human wisdom. On popular information platforms such as Baijiahao, we sometimes see some content of extremely poor quality, poorly made up sentences, almost gibberish, which are probably captured and spliced ​​by programs, not from Such “articles” cannot be regarded as works.
  With the advancement of technology and the maturity of users, the quality of content generated by AI is getting better and better, and the program can output certain original content under the leadership of the author. Such articles may be regarded as works under the copyright law . For example, the AI ​​painting “Space Opera House”, which won the first place in the Colorado State Fair art competition in September last year, was obtained after hundreds of adjustments by the author.
  Case 2: AI articles are works of legal persons
  One day in 2018, Dreamwriter, an intelligent writing assistance system developed by Tencent, created an article titled “Lunch Review: The Shanghai Stock Exchange Index rose slightly by 0.11% to 2671.93 points and communication operations, oil exploration and other sectors led the rise” The financial report, the article is based on the collection and analysis of a large number of financial articles on the stock market, the historical data of the stock market and the real-time stock market data of the day, and the unique expression wishes of the main creators. The article is completed and published within 2 minutes after the end of the stock market trading time on the day Note “This article was automatically written by the Tencent robot Dreamwriter”.

  On the same day, Shanghai Yingxun Technology Co., Ltd. (hereinafter referred to as “Yingxun Company”) copied the article and disseminated it to the public through its own website. Tencent believed that it had infringed the copyright and sued to the People’s Court of Nanshan District, Shenzhen.
  After the trial, the court held that the article involved in the case was generated by the main creative team of Tencent using Dreamwriter software, and its external performance met the formal requirements of written works, and its content reflected the selection, analysis, and judgment of relevant stock market information and data that morning. The structure of the article is reasonable, the logic of expression is clear, and it has a certain degree of originality. The court found that the article involved in the case was an overall intellectual creation formed by a multi-team and multi-person division of labor, which as a whole reflected Tencent’s needs and intentions for publishing stock review and summary articles, and was a legal person’s work presided over by Tencent.
  The court finally ruled that Yingxun provided the content of the allegedly infringing article to the public on its website without permission, for the public to obtain at a selected time and place, which infringed on the information network enjoyed by Tencent. The right of dissemination shall bear the corresponding civil liability. Since Yingxun has deleted the infringing works, the court judged the compensation amount to be 1,500 yuan.
  According to Article 9 of the Implementation Regulations of my country’s Copyright Law, the subjects who can enjoy copyright are citizens, legal persons or other organizations. Persons (including natural persons and legal fictitious persons such as legal persons and other organizations) are the only subjects who enjoy rights and perform obligations according to law. In this case, Tencent is not only the developer of the Dreamwriter program, but also the user of the program, that is, the generator of the content of the article. It has participated in the entire content production process and can enjoy the copyright as a legal person.
  The copyright ownership of AI works is nothing more than the program developer (platform side), program user (content generator), and the original author of the material, but usually only the first two are recognized, because there are too many materials used in the training program, it is difficult to determine so many The copyright of the original author. The really tricky problem is: if the developer and user of the program are not the same person, how to determine the copyright owner? The dispute is still raging.
  Not long ago, an individual author in the United States “created” a graphic novel called “Zaya at Dawn”. The pictures in the book were drawn by Midjourney software. The author entered the text content into the software, and AI made corresponding pictures. The signature on the cover of the book is the name of the person and the name of the software juxtaposed. In February this year, the Copyright Office of the United States made a ruling on the copyright dispute in this case, and adopted a demarcation method to limit the protection boundary of copyright. It believed that only the part written and arranged by the author can obtain copyright protection, and the part of the picture is not a human work. , are not protected by copyright. That is to say, only a part of the content of this work is protected by copyright, and this copyright is enjoyed by the human author.
  The copyright issue of AI works still needs more judicial practice to point out the direction. Zhou Bo, a senior judge of the Supreme People’s Court, once wrote that from a technical perspective, artificial intelligence has not yet developed to a level where it can truly get rid of human participation. If the products related to artificial intelligence are also the achievements of human intellectual activities assisted by artificial intelligence, then the existing legal norms are sufficient to adjust the copyright protection of products involving artificial intelligence. As for the copyright protection of artificial intelligence autonomous products without human participation at all, further observation is needed.

error: Content is protected !!