ARTIFICIAL INTELLIGENCE (AI) AND COPYRIGHT

It is common knowledge that that the law progresses more slowly than modern life. The law is not able to regulate absolutely all life situations, but only the most important ones. In order for some issues to be imposed as more important than others, the passage of time is necessary. This is the main cause of the phenomenon that legal regulations are lagging behind everyday life. However, the increasingly topical problem is that even everyday life is losing step with the progress of artificial intelligence.

What can Artificial Intelligence do?

In recent times, Artificial Intelligence – AI has achieved a significant presence in everyday life and work in the arts and creative sector. Some of the most popular AI systems in these areas are Amper Music, DALL-E and Chat GPT. Amper Music is an AI system that can create music based on set parameters. DALL-E is an AI system that uses so-called deep learning to create visual content based on the task that the user enters in the form of text. Its capabilities include creating artistic paintings in the style of famous painters, such as Van Gogh, Picasso and Monet.

Chat GPT, which has attracted the most media attention, can be applied in various fields, including customer support, sales and marketing, research and development, education and in many other fields where there is a need for effective and personalized communication with users.

So how do we identify AI-generated content?

The influx of new possibilities has certainly made it easier for people to perform everyday tasks, starting from composing emails to designing advertising campaigns and strategies, but it has also brought with it the use of AI for writing homework or term papers. However, it is important to keep in mind that these tools (especially free ones) have limitations in terms of the quality of the generated content and there is a need for further editing, in order to obtain satisfactory results.

Since artificial intelligence systems are constantly working to ensure that their automatically generated content is a mimicry of human-generated author content, the differences between these creations are becoming less and less noticeable. There are tools available to verify the authenticity of a work, but not all are equally effective in distinguishing between the creations of AI systems and works created by an author. For example, some software focuses on analysing textures and other visual elements to determine authenticity, while others use more sophisticated data analysis and machine learning techniques. There is an idea that works created by AI should embed a “watermark” invisible to humans, in order to prevent artificial intelligence systems from “learning” from earlier creations of artificial intelligence, which would make new works of AI programs look less and less like human-generated works.   

Advances in artificial intelligence and copyright

The increasingly rapid progress of artificial intelligence opens up many questions and challenges in all spheres of human activity, including law. The issue of copyright for AI created works is one of those that attracts a lot of attention from the public and legal experts, along with the inevitable and increasingly topical issue of personal data protection. The question arises as to who is a copyright holder when a program independently creates a work or when a human uses an AI framework to create a work. Although this issue is still not fully resolved in the legal system, there are some cases that have attracted the attention of the media and the public.

There are two schools of thought regarding the question of who owns the copyright on works created by AI. The first school of thought believes that the copyright holder is the creator of the AI system, since AI is just a human-made tool. The second school of thought believes that the copyright holder should be the AI program that created the work, because artificial intelligence has the ability to make decisions about the design of the work, therefore it can be considered to have some kind of creativity.

What does our law say?

The applicable  Law on Copyright and Related Rights  specifies a natural person as the holder of copyright on a work of authorship, with the exception of a film work, where a legal entity – the producer, whose name is indicated on the work itself, can also be considered the copyright holder. Our law recognizes certain rights to certain legal entities – producers of phonograms, videograms, broadcasts and databases; however, these rights are related rights, they are not copyright, and the subject of protection is also different.  

The Law on Copyright and Related Rights specifies which works are considered to be copyright. The legal definition of a work of authorship leaves no room for a broader interpretation, according to which an invention in the field of technology (which an AI program actually is) could be considered a work of authorship, since inventions in the field of technology are recognized as patent rights.

Bearing in mind the domestic regulation, the first school of thought fulfils the condition that the holder of the right is a natural person, but it is contrary to the nature of the copyright that the holder of the same is someone who only indirectly participated in the creation of the work, and that it was not created as a result of creative and intellectual engagement of the holder (provided that we recognize the creation of AI as a work of authorship).

The argument of the second school of thought that the copyright holder should be the AI program that created the work is legitimate if considered from the point of view that the copyright holder should be the entity that directly participated in the creation of the work. The disadvantage of this point of view is that the AI program as such has no legal subjectivity, therefore it cannot be the copyright holder, since it is neither a physical nor a legal entity.Even if this were ignored and the AI’s copyright were recognized, a number of problems would arise when exploiting it, since many artificial intelligence programs are capable of making rational business decisions.  

After all, our law stipulates that a work of authorship is the original intellectual creation of the author. The domestic doctrine, specialized in intellectual property rights, took the position that a necessary condition for a work to be considered work of authorship is that a person created it with their own work. In the textbook “Intellectual Property Law”, professors Slobodan Markovic and Dusan Popovic expressed the view that “works created by the activity of animals or devices are not creations in the sense of copyright”.

What does court practice say?

One of the most famous court proceedings related to this issue is the “Monkey selfie”, a dispute from 2015. A monkey took a selfie using photography equipment. Photographer David Slater, whose idea it was for the monkey to take a selfie and who set up the equipment for the venture, sold the photo, but PETA (an animal rights organisation) filed a lawsuit against him and the self-publishing company Blurb, claiming that the monkey should own the copyright to the photo.

After lengthy court proceedings, the court dismissed PETA’s lawsuit, arguing that animals cannot be copyright holders in the US. According to the appeal, a second-instance procedure was conducted, and the dispute was settled out-of-court. The U.S. District Court’s ruling did not set a precedent for extending the domain of copyright to animals, so it is questionable whether the same will be the case in disputes related to works of artificial intelligence.

A lot of public attention has been attracted by the dispute that is currently being held before the High Court of Justice in London, following a lawsuit filed by Getty images (a supplier of images) against Stability AI (an AI development company).

According to the plaintiff’s statement, Getty Images’ position is that Stability AI illegally copied and processed millions of copyrighted images and associated metadata owned or represented by Getty Images, without a license in favour of the commercial interests of Stability AI and to the detriment of content creators. According to the statement, Getty images has provided licenses to leading technological innovators for purposes related to the training of artificial intelligence systems in a manner that respects personal and intellectual property rights, and defendant Stability AI is not one of the licensed innovators.

It seems that there is a low degree of awareness about the importance of copyright and related rights in Serbia. This assertion is also supported by the absence of domestic court practice concerning these issues. The influence of artificial intelligence in the field of creative work will inevitably encourage consideration of the dilemmas presented. It is necessary to conduct a serious public discussion in Serbia on the protection of copyright and related rights in the era of artificial intelligence, in order to adapt the legal framework to modern technological trends. This is the only way to create a business environment that protects the rights of innovators.

Law Firm Petrovic Mojsic & Partners

Aleksandar Popović