Xinhua Wanghai, May 30 (Wang Ruoyu) 15 seconds to draw a movie poster, 1 minute to make a beautiful video, not only can write novels, poetry, copywriting, knock code, but also can make study plans, assist teaching, test questions. Recently, including ChatGPT, some language tool platforms driven by artificial intelligence technology have attracted attention from all walks of life. While people lament its "superpowers", it also raises corresponding risks and hidden dangers such as intellectual property rights, data security, and technical ethics, which urgently need to be paid attention to and paid attention to.
Recently, the "Series of Forum on Intellectual Property Protection in the Digital Era - Starting from the Intellectual Property Issues and Risks in the Use of ChatGPT" jointly sponsored by the Shanghai Association for Science and Technology, Shanghai Municipal Bureau of Market Regulation and East China University of Political Science and Law was held in Shanghai. After the meeting, Xinhuanet interviewed relevant scholars who participated in the event to answer questions of concern.
Xinhuanet: Will ChatGPT "grab jobs" with young people?
The picture shows Professor Wei Hui of the School of Computer Science of Fudan University. Xinhua Net (Photo provided by the organizer)
Professor Wei Hui of the School of Computer Science of Fudan University: What artificial intelligence can do is nothing more than reasoning, problem solving, decision-making, understanding, and learning. When solving practical problems, it mixes these five aspects. Behind this is still an algorithm. So AI won't really replace people's jobs, at least not at this stage. It can replace the work that can give solutions through machine learning, such as a junior secretary, writing something very formatted, just change a few words, change the combination of graphics, which is likely to be "laid off". But complex programming it does not do well, so real coders do not have to worry about this problem.
In addition, the so-called "superpowers" of artificial intelligence today are likely to be through the advancement of other technologies, such as computer networks, wireless communications, computer hardware, and similar to digitalization, big data models, and so on. These credits cannot all be attributed to artificial intelligence. For example, 5G to achieve unmanned driving, everyone will think that this is the progress of artificial intelligence, it is likely to use remote control technology, people are driving, not machines are driving.
Xinhuanet: How to manage the copyright risk of generative AI content?
The picture shows Cong Lixian, Dean of the Intellectual Property School of East China University of Political Science and Law. Xinhua Net (Photo provided by the organizer)
Cong Lixian, Dean of the School of Intellectual Property of East China University of Political Science and Law: Generative AI largely digitizes other people's works in data mining, and all learning is a variety of data presented by symbolic images from the perspective of semiotics. Massive mining uses works, or turns works into all symbols and images. This form of mining is very special and requires thinking about the copyright risks here.
I believe that the relevant problems can be solved from a technical point of view. For example, the introduction of watermarking technology, cryptography technology, labeling and other means, now many scientists have begun to try. In addition, it is also necessary to pay attention to platform copyright autonomy, and in the future, as more and more products are launched by various large companies, generative AI will become a widely used Internet iterative technology. Platform autonomy and cooperative governance between platforms on copyright issues had become very important.
Xinhuanet: Are the works produced by artificial intelligence inventions? Can AI become the subject of intellectual property?
The picture shows Professor Xu Chunming of Shanghai International Intellectual Property Institute of Tongji University. Xinhua Net (Photo provided by the organizer)
Professor Xu Chunming, Shanghai International Intellectual Property Institute, Tongji University: According to actual laws, works protected by the Copyright Law must be the creative achievements of human beings and the expression of the author's thoughts and emotions; The invention-creation in patent law must be the invention and design of human beings, and the intellectual achievement of the inventor. Artificial intelligence products are not the creation and creation of natural people, and do not reflect human thoughts, emotions and intellectual creation. Therefore, the current position of the legal profession is: AI creations are not works, AI cannot be authors, and AI creations are not protected by copyright; AI creations are not inventions, AI cannot be inventors, and AI creations are not patentable.
In terms of legal application, AI products are not protected by copyright law and patent law, which does not mean that they are not protected by other laws. AI products are not human intellectual achievements, but can be identified as human labor achievements, are the labor results produced by generative AI owners based on certain inputs or even huge inputs, have commercial value, can form market competitive advantages, have protectable legal benefits, and can be protected by applying the Anti-Unfair Competition Law to achieve copyright incentives for the development and application of generative AI.
Xinhuanet: What new challenges does the era of artificial intelligence bring to intellectual property protection?
The picture shows Liu Feng, director of the Intellectual Property Business Committee of the Shanghai Lawyers Association. Xinhua Net (Photo provided by the organizer)
Liu Feng, Director of the Intellectual Property Business Committee of the Shanghai Lawyers Association: The working principle and premise of artificial intelligence is to widely capture various data. At this point, the first level of questions arises: will scraping infringe on others' trade secrets? Does scraping involve data security? Secondly, we will also encounter a second level of challenges, ChatGPT does not simply copy and copy after capturing data, but processes the data, and even complies with the requirements of the Copyright Law to re-edit the text and avoid the comparison of "checking duplicates". So, the question is, how to protect the rights of the original copyright owner has encountered a problem, how to prove that the artificial intelligence has captured its own work and then modified and edited? These problems will soon be put in front of enterprises and intellectual property rights holders. Scientific and technological personnel and legal workers should fully explore these frontier issues, which is of great significance for the revision and formulation of mainland laws in the future.