Copyright of objects created by artificial intelligence (AI) is becoming an increasingly relevant topic in the context of rapid technological development. Since the emergence of generative models such as ChatGPT, Midjourney, and others, there are many legal issues regarding the ownership of copyrights for using the results of their interaction by the user. In particular, who is the author of a work created by AI; can objects created by AI be protected by copyright; who is responsible for violating the rights of third parties if AI uses data protected by copyright, etc.

First of all, it should be noted that objects created with the help of artificial intelligence (AI) have a number of specific features that seriously affect their legal status and the issue of determining authorship, which in turn gives rise to the problem of establishing the owner of property rights to such objects, namely:

• Use of existing information and data. AI works by analyzing huge amounts of data that can include text, images, videos, and other materials that are also protected by copyright. In addition, it is important to understand that the results generated by AI are often a compilation or adaptation of this existing data. For example, AI can create a new image by mixing parts of several previously created paintings, but without introducing any new creative element, which subsequently raises questions about the legality of their use and possible copyright infringements by authors.
• Lack of a creative approach in works created with the help of artificial intelligence is one of the key factors affecting their legal status. In a legal context, for a work to be considered a copyrighted work, it must have an element of creativity, that is, be the result of human intellectual activity. This creative approach includes uniqueness and originality, which are the result of the author's (creator's) personal, subjective decisions. At the same time, any generative AI model acts as a technological tool that operates based on programmed algorithms that process data and generate new content based on specified parameters. However, despite the complexity and novelty of the created works, the model acts automatically, without personal creative intention or conscious choice. This means that the results of the work are only a mechanical combination of data, without the participation of personal judgments, imagination, or emotions inherent in human creativity.
• Generation of works without human participation. First of all, there is a fundamental question of whether such works can be considered the result of intellectual activity at all. In particular, if AI generates a work automatically, without human intervention, the question arises: who is the author of such a work? According to the classical approach of copyright, the author is a natural person who creates a work, that is, one who makes an individual creative contribution. In the case of works created by AI, the lack of human participation in the process of creating a work means that there is no specific person who can be recognized as the author, which creates a legal vacuum where the object (work) exists without clear ownership of copyrights, and therefore there is a lack of framework for determining property rights to the object. At the same time, even taking into account the peculiarities of such objects and the potential impossibility of protecting such objects within the framework of copyright, the solution to this issue varies significantly in different countries, creating a number of precedents and different approaches to dispute resolution practices.

United States

In the United States, copyright is granted only to original copyrighted works fixed in any tangible medium, so products fully created by AI cannot receive copyright protection. However, the impetus for resolving this issue was the case of Christina Kashtanova regarding the comic "Zarya Of the Dawn" created by her using artificial intelligence, where it was specifically determined to grant copyright only to the comic as a whole (collection), but the individual images generated by Midjourney will not be subject to protection, since "Kashtanova's actions in this case were too minor and inconspicuous... to provide a sufficient level of creativity for copyright protection of individual images". As a result of this case, in May 2023, the US Copyright Office issued "Instructions for Registering Copyrights in Works Containing Material Generated by Artificial Intelligence", which clearly stated that "individuals who use artificial intelligence technology to create a work (object) may claim copyright protection for their own contribution to that work (object) by identifying in the standard application the portion of the work (object) to which the author claims rights".

In addition, a number of lawsuits against GitHub Copilot (an AI coding assistant created on pirated open source software), Stability AI, DeviantArt and Midjourney, Runway AI, as well as Google (Imagen tool) from Sarah Andersen, Kelly McKernan, Jingtao Zhang, Grzegorz Rutkowski and others for copyright infringement contributed to the creation and presentation to the general public in July 2024 of a project by the US Copyright Office to resolve the issue of copyrights for artificial intelligence objects in order to create a unified approach to identifying problems, establishing terms for identifying AI objects, as well as ways to resolve copyright infringements within the framework of using artificial intelligence tools.

European Union

In European Union, regulation of objects created by AI is in an active development stage, especially with the introduction of the "AI Act." This project aims to regulate the use of AI, including issues related to copyright. The main focus is on assessing the risks to human rights and freedoms arising from the use of AI. At the same time, the issue of whether objects created by artificial intelligence belong to "works" was resolved within the framework of the EU's final report on artificial intelligence trends and development in 2020, where it was determined that an object created by AI will be considered a "work" if it meets the following criteria: 1) belong to the literary, scientific or artistic field; 2) be the result of human intellectual work; 3) reflect the creative choice of a person, which manifests itself in this object.

Ukraine

In Ukraine, the situation regarding the regulation of copyright for AI-generated objects remains less defined. The legislation does not contain specific provisions concerning objects created by AI. However, the issue has been partially addressed within the framework of the Law of Ukraine on Copyright and Related Rights under a special right - sui generis. In this case, the concept of "non-original objects generated by a computer program" arises, which are not recognized as works in their original definition, since a work is an original intellectual creation of the author. Nevertheless, the creation of such an object requires the involvement of human resources to achieve a specific result (product), and accordingly, there is a protection of property rights (permission or prohibition of the use of the object) in order to gain benefits from the engaged human resources in creating such an object. Thus, the application of sui generis rights to objects created by artificial intelligence is based on the premise that the criterion for obtaining protection is not originality, as in the case of copyright, but the form of expression (AI work result) and human resources involvement, where AI is defined as a tool through which the object was created. At the same time, the subjects of sui generis will be considered individuals who hold property rights to such an object and which need to be protected. Specifically, the sui generis right arises from the moment of the creation of a non-original object (an object created by AI), and the term of validity of property rights is 25 years.

United Kingdom

According to the legislation of the United Kingdom, specifically the Copyright, Designs and Patents Act 1988, artistic works created using computer programs can be protected by copyright (although there is no copyright registration system in the UK). In particular, Section 178 of the Act states that a "work created by a computer program" is a work created by a computer without human involvement, and the "author" of the computer-generated work will be considered the person who took the necessary steps to create the object. Such works will be protected from the moment of creation and for 50 years from the end of the calendar year in which they were created.

China

In China, copyright law does not provide clear provisions for regulating the issue of objects created using artificial intelligence. Specifically, to ensure copyright registration and obtain a certificate for a work in China, the object must meet two conditions: (1) it must be an original intellectual creation in the fields of literature, art, or science; and (2) it must be capable of being presented in a specific form. Based on these positions, a hearing took place in the case of Li vs. Liu (Jing 0491 Min Chu No. 11279), where the plaintiff Li created several images using input prompts in Stable Diffusion (an open-source generative AI model) titled "The Spring Breeze Brought Tenderness," and the defendant Liu, using the current image, removed the plaintiff's signature and posted the object on a popular creative content exchange platform. Thus, the question arose regarding the protection of the created object under copyright and whether there was an infringement of the plaintiff's copyright.

In this case, several issues regarding the positioning of copyright for AI-generated objects were immediately resolved: 1) it was determined that an image that does not externally differ from an ordinary photograph or painting belongs to the realm of art, can be presented in a specific form, and qualifies as a work of visual art; 2) the court established that through independent design and improvement of visual elements and composition using several rounds of input prompts and parameter adjustments, as well as through the artistic choice of a desired close-up portrait of a woman in twilight, the plaintiff participated in the process of "intellectual creativity," resulting in an "original" work rather than a mechanical creation; 3) overall, copyright for an object created in China is held by its author, who can be a natural person, legal entity, or incorporated organization. The court stated that the AI model itself cannot be an author, and the developer of the AI model did not participate in creating the image (he acted solely as the producer of a creative tool). Moreover, the user license explicitly states that Stability AI does not claim or retain any rights to the output content generated by the tool. Thus, the court ruled that Li, as the author of the image, is the legal owner of the copyright for the generated image.