Turkey: Analysing the relationship between AI and copyright law
Artificial intelligence (AI), which does not yet have a solid definition, can generally be described as “the ability of a digital computer or computer controlled robot to perform tasks commonly associated with intelligent beings.” AI systems have started to play a part in our daily life and today they carry out crucial roles for human beings. For instance, BlueDot’s AI system was among the first in the world to identify the emerging risk from COVID-19.
AI also creates music, art and, furthermore, today, AI can write poems, novels, stories, and legal articles. The Next Rembrandt is one of the outstanding examples of an AI-generated work of art. With the supporting partnership of ING Bank, Microsoft, advertising agency, J Walter Thompson, and advisors from Delft University of Technology (TU Delft), as well as the Mauritshuis and Museum Het Rembrandthuis, we had the opportunity to see what kind of painting Rembrandt Harmenszoon van Rijn, the Dutch painter, could create if he were alive, today
As a result of these developments, discussions have begun regarding AI-created works of art and the notion of ownership and enforcement, within the context of copyright law.
In modern society, scholars ask whether an AI-created work can be copyrighted, and whether the AI can be regarded as the owner of a copyright. If AI is considered the author of a work of art, the doctrine examines the problems that the courts might face, given the current legal framework of copyright law. For example, the issue of accountability has become a frequently raised question in the event of infringement by AI systems because, in the event of an infringement, what benefits can copyright holders derive from suing an AI system, and who would be the infringing party in this scenario? Since the plaintiff most likely would not receive any compensation from the AI, considering the fact that AI does not have any assets, this could neutralise the effect of the violation. Most importantly, what form of deterrence might a law have on AI? Moreover, the enforcement of copyright law by AI may be problematic, as well, because the present rules do not answer how, and in what way, an AI system can enforce copyright, who might act on behalf of AI, and how the legal representative may be appointed.
Even though case law on this subject is too immature to have come to a conclusion, it is important to note recent developments. In the United States and Australia, judges did not recognise AI as the creator, and ruled to afford protection only to works created by humans. However, on the other side of the world, a Chinese court decided that neither the software developer, nor the user, is the author and, after analysing whether a software may be regarded as the owner of the work in question, the court ruled that the work cannot be copyrighted. Nevertheless, considering the composition of the work, and the input of the software developer and the user, some sort of protection should be afforded. Currently, a similar issue is being assessed by a Canadian court. Considering the lack of legal framework surrounding AI-created works, it is evident that the courts are constrained by the current rules, which do not provide enough space either for ground-breaking judgments or for interpretations.
On the other hand, in the Resolution of the International Association for the Protection of Intellectual Property (AIPPI) on Copyright in Artificially Generated Works dated September 18 2019 in which 30 countries took part, besides other issues, AIPPI resolved that AI generated works should only be eligible for copyright protection if there is human intervention in the creation of the work and provided that the other conditions for protection are met. Accordingly, this resolution raised the necessity of flexing the strict approach in current copyright rules, which do not provide protection to AI-generated works, and suggested enabling it at least for the works where human intervention is involved.
Situation in Turkey
Turkey’s current copyright law does require a real person to be recognised as author of a copyrightable work, and therefore it is not possible for AI to be the author and/or enjoy copyrights.
While we have seen no Turkish case law concerning AI-created works and copyright yet, this subject has become ubiquitous among Turkish scholars. There are many articles that discuss whether AI should be regarded as the author of a work of art, and what the benefits and problems are associated with it. At the same time, some Turkish scholars study the subject from another perspective and suggest the use of work-made-for-hire rules. Until copyright rules are modernised, they suggest using work-made-for-hire rules to overcome the challenges of ownership, enforcement, or accountability regarding AI-created work within the context of copyright law.
The work-made-for-hire doctrine is an exception to the general principle of copyright ownership by giving employers or the individuals commissioning the work, the copyright in works of authorship created by employees or subcontractors. Accordingly, some Turkish scholars suggest that the relationship between AI and its owners/developers is similar to the relationship between employer and employee. It would therefore be possible to apply the work-made-for-hire doctrine to AI-generated works and enable the developer/owner to hold the authorship of these works. However this approach has been criticised by others due to its lack of applicability. Firstly, Turkish law does not apply the work-made-for-hire rules. Secondly and more importantly, this doctrine requires real people as employees or individuals commission their works, which would clearly not be the AI itself since it is neither recognised as a real person nor a legal entity yet.
There are also other suggestions in line with discussions all around the world, such as attributing the status of a legal entity to AI and enabling this entity to hold rights and to have responsibilities arising from the AI’s acts, introducing ethical rules for the whole process of AI-generated works, while providing rights to people/companies behind the AI, such as owners and programmers, attributing an electronic personality to AI or, finally, placing the AI-generated work into the public domain.
Even though, currently, the case law and the doctrine do not clarify legal uncertainty surrounding AI-created works within the context of copyright law, considering the constant technological advancements in the field of AI, we believe that this topic will continue to be a point of discussion in the upcoming years.
Mutlu Yildirim Köse and Havva Yildiz