
ChatGPT, a language model, aims to create meaningful content in line with the commands entered. The legal nature of the content produced by ChatGPT, which is mostly successful in producing meaningful content, is debatable. In this context, the content in question should be considered in terms of intellectual property law, or more clearly, the content produced by ChatGPT should be examined in terms of copyright rights. It should be noted that since copyright rights in Turkish Law will only apply to works, the first thing to be examined is whether the content in question has the quality of a work.
According to Article 1/B of Law No. 5846; All kinds of intellectual and artistic products that bear the identity of their owner and are considered as works of science, literature, music, fine arts or cinema have the quality of a work.
As can be understood from the aforementioned definition, in order for a product to be considered an intellectual and artistic work under Law No. 5846, it must have four characteristics.[1] First, it must be an intellectual and artistic work. Second, the work in question must bear the characteristics of its owner. Third, it must be a formed product that has not remained at the intellectual stage. Finally, the product must fall within the scope of one of the types of works included in Law No. 5846.
First, it is necessary to mention whether the works produced by ChatGPT are intellectual and artistic works. There is no definition of what an intellectual and artistic work is in Law No. 5846. However, it is accepted in the doctrine that only things produced by humans can be considered intellectual and artistic works. Therefore, human influence should be considered in the content produced by ChatGPT. In the content produced by ChatGPT, the human influence is only entering a command consisting of a few words or sentences. Even if this is the case, we believe that in the analysis of human influence in the content produced, it is not the length of the text entered by the human, but the final influence of the real person in presenting the text that is important. In other words, if the command entered by the real person had not existed, the content in question would not have been produced by ChatGPT. Therefore, since the effect of the real person in the production of the content is obvious, it is possible to say that the works produced by ChatGPT have the characteristics of intellectual and artistic products.
Secondly, the work in question must bear the identity of its owner. There is no definition of identity in Law No. 5846. The General Assembly of the Court of Cassation, in its Merits No. 2017/11-2758, Decision No. 2021/1228 and dated 14.10.2021, briefly defined identity as “the characteristics that the owner of the work adds to his work from his own world of ideas, thoughts, emotions and feelings and his abilities based on intellectual foundations”.[2]
Although the author is defined as the person who creates the work in Law No. 5846, both the doctrine and the Supreme Court of Appeals are of the opinion that the author can only be a real person. Therefore, ChatGPT, which does not have a real personality in terms of law, cannot be considered as the author in terms of the content it produces. In this case, it would be appropriate to examine the nature of the real person who enters the command for ChatGPT to produce content. In most cases, the person who enters the command enables ChatGPT to produce relatively much longer content with a few words or sentences. It is also observed that ChatGPT gives different answers to the same command. Therefore, although it is a system that initiates the content production process with human influence, it cannot be claimed that the human factor is the source of originality in the production of content. In other words, although ChatGPT produces content according to the command it receives, the content in question does not have an original quality for this reason alone. In fact, the person giving the command can obtain different content with the same command. For these reasons, the nature of the real person who enters the command cannot be mentioned on the work.
As a result, the content produced by ChatGPT has the other three characteristics in the definition of a work of art defined in Law No. 5846, but does not meet the requirement of having the author’s identity. Since ChatGPT cannot be considered a natural person, it cannot meet the requirement of having the author’s identity, while the user who enters the command cannot provide original content production due to different results with the same command and therefore does not meet the requirement of identity. Therefore, it is not possible for the content produced by ChatGPT to be accepted as an intellectual and artistic work according to Law No. 5846.
We find it appropriate to end the discussion with the sentence in ChatGPT’s answer to the question “What would be the last sentence of an article stating that ChatGPT cannot replace human creativity?”: “The source of creativity is humans, no machine can imitate it.”[3]
REFERENCE
GÜÇLÜTÜRK, Ö. Ü. O. G. (2022). CHATGPT İLE ÜRETİLEN İÇERİKLERİN ESER NİTELİĞİNİN 5846 SAYILI FİKİR VE SANAT ESERLERİ KANUNU BAKIMINDAN DEĞERLENDİRİLMESİ. Galatasaray Üniversitesi Hukuk Fakültesi Dergisi, (2).
Law No. 5846 on Intellectual and Artistic Works and Related Legislation
[1] GÜÇLÜTÜRK, Ö. Ü. O. G. (2022). CHATGPT İLE ÜRETİLEN İÇERİKLERİN ESER NİTELİĞİNİN 5846 SAYILI FİKİR VE SANAT ESERLERİ KANUNU BAKIMINDAN DEĞERLENDİRİLMESİ. Galatasaray Üniversitesi Hukuk Fakültesi Dergisi, (2).
[2] GÜÇLÜTÜRK, Ö. Ü. O. G. (2022). CHATGPT İLE ÜRETİLEN İÇERİKLERİN ESER NİTELİĞİNİN 5846 SAYILI FİKİR VE SANAT ESERLERİ KANUNU BAKIMINDAN DEĞERLENDİRİLMESİ. Galatasaray Üniversitesi Hukuk Fakültesi Dergisi, (2).
[3] GÜÇLÜTÜRK, Ö. Ü. O. G. (2022). CHATGPT İLE ÜRETİLEN İÇERİKLERİN ESER NİTELİĞİNİN 5846 SAYILI FİKİR VE SANAT ESERLERİ KANUNU BAKIMINDAN DEĞERLENDİRİLMESİ. Galatasaray Üniversitesi Hukuk Fakültesi Dergisi, (2).