УДК 347.78

Civil liability for copyright infringement using artificial intelligence technologies

Кутузова Екатерина Николаевна – студент магистратуры Юридического института Владимирского государственного университета А.Г. и Н.Г. Столетовых.

Научный руководитель Новикова Людмила Васильевна – кандидат филологических наук кафедры иностранных языков профессионального общения Владимирского государственного университета А.Г. и Н.Г. Столетовых.

Abstract: In connection with the development of artificial intelligence technologies, legal scholars and legal practitioners are faced with the question of copyright protection, since the very existence of artificial intelligence makes copyright subjects vulnerable. In this article, the author examines the reality of civil liability for copyright infringement using artificial intelligence and also proposes measures to improve copyright legislation, taking into account the capabilities of artificial intelligence to prevent possible violations.

Аннотация: В связи с развитием технологий искусственного интеллекта перед правоведами и юристами-практиками встает вопрос о защите авторских прав, поскольку само существование ИИ делает уязвимыми субъектов авторских прав. В данной статье автор изучает вопрос о реальности гражданско-правовой ответственности за нарушение авторских прав с использованием ИИ, а также предлагает меры по усовершенствованию законодательства в области авторского права с учетом возможностей искусственного интеллекта для предотвращения возможных нарушений.

Keywords: artificial intelligence, neural network, copyright, exclusive rights.

Ключевые слова: искусственный интеллект, ИИ, нейросеть, авторское право, исключительные права.

The rapid development of science and technology in the 21st century not only has positive aspects, but also opens up opportunities for unscrupulous people who are finding new ways to commit crimes. Recently, in the wake of the popularity of artificial intelligence (AI), methods have emerged for its use both for the benefit of humanity, facilitating many processes in various areas, but also as a new means of committing crimes. In this regard, the relevance of this topic cannot be underestimated. Legal scholars have just begun to turn their professional attention to artificial intelligence technologies in the legal context. And despite their attempts to determine its role in legal relations, there is still a lack of understanding of AI as a phenomenon in law and its role in the system of legal relations. And the lack of full consolidation of its status in the law makes it possible for unscrupulous participants in legal relations using this implement to take advantage of the vulnerability of other legal subjects.

Of course, legal gaps are inevitable. Law and law only follow the development of social relations and innovations in the field of science and technology. It is in the interests of the whole society that legal scholars and legislators eliminate emerging legal gaps and conflicts as soon as possible.

So, as is known from the theory of law, all offenses are divided into: crimes - guilty socially dangerous acts prohibited by criminal law, and misdemeanors, which in turn are classified into civil, administrative, and disciplinary. In this study, we will turn to civil law relations and offenses in the relevant branch of law, since copyright is a subfield of civil law.

The legislator has established provisions related to the protection of copyright in Part IV of the Civil Code of the Russian Federation in Section VII of this Code [1], which discloses the rights of the author to the results of his intellectual activity and means of individualization. Copyright protection by law presupposes, according to paragraph 2 of Art. 1255 of the Civil Code of the Russian Federation: the exclusive right to a work, the right of authorship, the author’s right to a name, the right to the inviolability of a work and the right to publish a work.

To determine how AI can disrupt the author's right to his work is worth describing the principle of operation of the neural network. Today there are several types of neural networks. The type of greatest interest for this study is generative and generative adversarial neural networks, which can create images, music, texts, etc. At first glance, it seems that the generated objects are unique, and this is only partly true. However, if you delve deeper into the process of creating these objects, then everything becomes ambiguous.

At the moment, the main difference between AI and human intelligence is lack of independence. Without a human-defined task, AI will not function productively enough. Having received a request in the form of a task, the AI begins to generate an answer based on existing data. These initial data become works created by other people, who may, among other things, be protected by law. At the moment, there are a number of neural networks that can fulfill a request to generate an image (Midjourney), text (ChatGPT) or a musical audio track, where the basis is the voice of a real person with his unique timbre, diction, natural vocal abilities (SoftVC VITS Singing Voice Conversion). The author gives examples of the most popular neural networks, but at the moment there are more and more of them, and not all of them learn from initial data from open sources. In some cases, permission from copyright holders may be required, which is usually not available. This issue is especially acute when creating content using AI for commercial purposes, because in this case the original author, who found similarities with his works, can claim compensation.

In this regard, humanity immediately faces several questions: the first is ethical, especially if neural networks use images (video) or voices (audio) of real people, and especially those individuals who are no longer alive. And the second, who follows from the first, is legal. As the Latin proverb says: Jus est ars boni et aequi (Law is the art of goodness and justice). In this connection, I would like to ask the question: is it fair to the authors of the work that AI users create new, but not inherently unique works with the help of their results of intellectual activity and consider they the authors of the result obtained? Is it even possible to apply the provisions of Section 4 of the Civil Code of the Russian Federation to the results obtained?

So far the legislator has not answered the last question as clearly as we would like. However, slowly but surely, legal scholars are inclined to believe that the user of a neural network does not have copyright on the results generated using AI.

If we turn to the letter of the law, we see a completely unambiguous answer to the question of who can be the author in the broad sense of the word in paragraph 1 of Art. 1228 of the Civil Code of the Russian Federation: the author of the result of intellectual activity is the citizen whose creative work created such a result. There is no talk of using AI technologies. However, it is worth considering that when this article was created, such a phenomenon as a neural network did not exist, that is, in theory, the interpretation can be expanded over time to the point that the category of creative work will include the use of AI as a tool.

However, if you look at the problem from a different angle, you can rightly ask the question: on what basis can an AI user not be the author of the result of his undoubtedly creative activity, provided that his personal contribution to the work was significant and the technology was used only at some stages of the creative activities? Perhaps, for law enforcers and legislators, the decisive role will be played by the percentage of contribution to the result of intellectual activity of AI technologies and the actual human intellectual activity, talent and skill.

Observing the development of AI, we see that biometric personal data, which is protected by the Federal Law «On Personal Data» [2] is under threat. Biometric personal data includes, according to paragraph 1 of Art. 11 of this Federal Law, information that characterizes the physiological and biological characteristics of a person, on the basis of which his identity can be established. There is reason to believe that voice is one of them. At the moment, there are AI technologies that can synthesize spoken speech of any content, where the voice of a real person is taken as a basis. However, at the moment there is no protection mechanism, for example, for professional announcers and dubbing actors, who, as practice has shown, become victims of neural networks that use voices without the consent of their owners. The Union of Russian Announcers spoke about this problem for the first time.

This is how the relevant disputes first appeared in court. For example, in the proceedings of the Savyolovsky District Court of Moscow, a statement of claim dated August 30, 2023, by dubbing actress A.S. Andronova was registered to Joint-Stock Company Tinkoff Bank on the protection of honor, dignity and business reputation. The parties entered into an agreement, the subject of which is the voicing of the customer’s texts for training the neural network and, according to the plaintiff, the result of the work was to be used exclusively for the company’s internal tasks. However, her voice turned out to be available for synthesis to everyone and is sold on third-party sites without her consent. The plaintiff demands that her synthesized voice be removed from all third-party sites, that her voice be stopped from being used in speech synthesis technologies available to third parties, and that she be compensated. The defendant's position is that the arguments set out in the claim are not true. The agreement expressly provides for consent to the use and processing of recordings by third parties. Therefore, the bank has the right to use the actress’s voice recordings in any way.

Since the court has not considered the case at the moment (11/18/2023), we can only guess what its position will be. But lawyers from the field of intellectual property law, discussing similar cases, believe that at the moment the right to vote cannot be protected by the rules of intellectual property rights because the list of IPR objects is exhaustive (clause 1 of Article 1225 of the Civil Code of the Russian Federation), and the human voice is not a type of personal non-property rights not related to property rights. At least judicial practice is limited to the list contained in paragraph 1 of Art. 150 Civil Code of the Russian Federation. For this reason, rules on the protection of personal non-property rights not related to property rights cannot be applied. The list of objects of protection in Art. 152.1 of the Civil Code of the Russian Federation is limited only to the image of a citizen [3].

Summarizing the above, we see that Russian legislation is not yet ready for the development of AI. Since there is still no mention of AI in regulatory legal acts, which has led to the fact that at the moment not only copyrights to the results of intellectual activity, but also biometric personal data of people are under threat. Without a new law “On the Use of Neural Networks” or amendments to civil legislation, it is impossible to create an orderly judicial practice, which will time after time lead to persistent violation of copyright on the results of intellectual activity and means of individualization.

The author of the study came to the conclusion that with the inevitable soon rule-making activity caused by the rapid development of AI and its implementation, including in commercial processes and business, the legislator should not artificially slow down the development of this technology with categorical restrictions for AI users, but create legal framework for its legal use for both ordinary and commercial purposes.

Literature

Civil Code of the Russian Federation (part four) dated December 18, 2006 No. 230-FZ (as amended on June 11, 2021) // Collection of legislation of the Russian Federation. – 2006. - No. 52 (1 part) - Art. 5496.
Federal Law of July 27, 2006 No. 152-FZ “On Personal Data” // “Rossiyskaya Gazeta” of July 29, 2006 (No. 165)
Mansurov, G. Z. Neural network and the right to vote / G. Z. Mansurov. — Text: electronic // Ural State Economic University: [website]. — URL: https://civilrecht.usue.ru/sobytiya/1131-nejroset-i-pravo-na-golos/ (access date: 11/19/2023).