In October 2025, during the "AI Driven Education Research Paper Writing" competition at East China Normal University, it was required that the submitted papers be mainly created by AI systems, with AI being listed as the first author and human researchers as co authors or corresponding authors, sparking discussions in the academic community. This article attempts to explore whether AI can serve as the first author from the perspective of intellectual property. The intellectual property right of authorship can be divided into two categories: one represented by patents, trademarks, and designs, and the other represented by copyright. The discussion in this article mainly focuses on the category of copyright. Currently, the academic community has a relatively consistent definition of "first author" and "corresponding author", which means that the first author is the main executor of the research, responsible for core tasks such as experimental operations, data collection and analysis, and writing the initial draft of the paper; The corresponding author is responsible for manuscript submission, revision, and communication with the journal, and assumes academic responsibility for the paper. Becoming an author means automatically enjoying copyright and personal rights. The former refers to the right enjoyed by copyright owners to use, profit from, and dispose of their works, thereby obtaining economic benefits, such as the right to copy, perform, adapt, and disseminate, which is transferable. The latter refers to the right enjoyed by the copyright owner of their work, which is closely related to their personal identity and has no direct property content, in order to maintain their personal connection with the work and the integrity of the work. It is enjoyed by the author for life and is generally non transferable, non deprivation, and non restrictive. Article 10 of China's Copyright Law stipulates four personal rights of authors: the right of publication, the right of attribution, the right of modification, and the right to protect the integrity of the work. Therefore, the scope of this article's discussion is refined to the "right of attribution" in the copyright of the author. AI is not the "author" in copyright law. Currently, in the copyright practices of major countries and regions such as China, the United States, Japan, the European Union, and South Korea, it is clearly stipulated that copyright can only be granted to natural persons or legal entities. For example, Article 11 of China's Copyright Law stipulates that copyright can only be granted to natural persons or legal entities. The EU Copyright Directive (Directive 2001/29/EC) stipulates that the creators of copyright are usually natural persons, and legal persons can become the holders of copyright. According to Section 201 of Title 17 of the U.S. Code, the holder of a copyright can be a natural or legal person. Article 7 of the German Copyright Law (Urheberechtsgesetz) stipulates that copyright can only belong to natural persons, but legal persons can act as owners of copyright. Article 11 of Japan's Copyright Law stipulates that the copyright belongs to the natural person who creates the work. In the case of job creation or contract creation, legal persons (such as companies, research institutions, etc.) can become the holder of the copyright. Article 10 of the Korean Copyright Law stipulates that the copyright owner is a natural person who creates a work. Currently, according to the universally accepted legal framework, AI is not an eligible subject for copyright. In addition to written law, a similar attitude is also upheld in the precedents of the Anglo American legal system. For example, in the case of Thaler v. Perlmutter (D.C. Cir. 2025) decided by the United States Court of Appeals for the District of Columbia on March 18, 2025, the plaintiff was computer scientist Stephen Thaler, the developer of the AI system Creativity Machine, and the defendant was Shira Perlmutter, the director of the United States Copyright Office. The plaintiff Thaler attempted to list the AI as the copyright "author" of the work, but the Copyright Office refused to list AI as the copyright "author", believing that the work must be completed by a human creator. The plaintiff filed a lawsuit to the court after the administrative reconsideration failed. After the first instance and appeal, the appellate court made a judgment, upholding the decision of the Copyright Office and confirming that the "author" in the Copyright Law must be a natural person, and AI cannot be the "author" of copyright. The judgment emphasizes that human creation is the core requirement for copyright registration, and based on the interpretation and legislative purpose of copyright law, the court believes that the "author" should be a natural person with creative intent and ability. The legal philosophy considerations behind consensus, whether in the civil law system or the common law system, seem to have reached a consensus on the idea that the author should be a natural person or a legal person, based on the "AI tool theory" of legal philosophy considerations. The fundamental purpose of copyright law is to protect the labor achievements of creators, especially their intellectual and spiritual labor. AI is currently widely regarded as a creative tool, similar to other technological means such as paintbrushes, computer software, or music synthesizers. AI tools can assist creators in generating text, music, or images that generate copyright, but ultimately the decision and control of the creation are still exercised by the creators. Therefore, as an automation tool, AI is not considered to have independent creative intent and personal rights attributes, and cannot become a copyright owner. Technically speaking, the Large Language Model (LLM) widely used for creation is a computer-readable tokenization of corpora (such as text, images, sound, images), followed by spatial vector localization. According to the demand instructions, AI provides results in an analogical manner, which is somewhat similar to "drawing a gourd and a ladle". The actual process of AI participating in academic work, such as generating a certain part of a paper, requires scholars to give instructions and identify the results produced by AI - whether there is real evidence, whether there is illusion, and to cut and select combinations of the result content. Natural persons still bear the key functions of decision-making, guidance, and control. The establishment of copyright by law not only protects the results of creative labor, but also clearly defines the responsibilities and obligations of copyright owners. If AI can act as a copyright owner, there may be issues with responsibility. For example, if an AI created work infringes on someone else's copyright or involves malicious or discriminatory content, there needs to be a truly responsible entity to take responsibility. There have been precedents in this regard. For example, on November 4, 2025, the High Court of England and Wales in the United Kingdom ruled in the case of Getty Images v. Stability AI, which involved the unauthorized use of millions of copyrighted images by the Stability AI model to train the model. In the images generated by the subsequent model, there appeared Getty company watermarks and trademarks with confusing features. In this regard, the British court held that the AI model of Stable Diffusion was trained with Getty images to form weight parameters for the spatial vectors of internal tokens, but it neither stored nor copied Getty images, thus not constituting infringement in "infusing copies" or "articles". From the perspective of intellectual property, it can be seen that in order for AI to become a copyright owner, the issue of subject responsibility and accountability must be considered, which is difficult to break through under the current framework. Facing the intellectual property challenges of AI creation, the author believes that although some of the settings of the essay competition at East China Normal University are contrary to the current legal framework, such an attempt is still meaningful. Currently, as AI's involvement in academic creation continues to deepen, it is necessary to have a clear understanding of its capabilities and presentation status. This essay contest is a great opportunity to explore the true capabilities of AI, allowing scholars to showcase their creative abilities and boundaries without hesitation, and helping the academic community to enhance their objective understanding of AI capabilities and shortcomings. With the continuous development of AI technology, the gradual improvement of intellectual property frameworks is an inevitable trend. For example, the European Commission raised ethical and legal issues related to AI technology in its 2020 AI White Paper, emphasizing the need to update relevant laws to adapt to technological developments. The white paper discusses the potential impact of AI on intellectual property rights such as copyright and patents. Although it does not change the basic principle that "creators are natural persons", the European Commission believes that existing laws need to be updated and further clarify the issue of copyright ownership of AI generated works. In short, when constructing intellectual property rules related to AI copyright in the future, ethical discussions should be emphasized, social impacts should be comprehensively considered, the auxiliary role of AI in the creative process should be clarified, and the transparency and fairness of academic research and evaluation should be ensured. (Liao Xinshe) Author: Wang Lei (Senior Intellectual Property Engineer of the Intellectual Property Development Research Center of the China National Intellectual Property Administration), Dai Ni (Class III Patent Examiner of the Patent Office of the China National Intellectual Property Administration)