1. Introduction
The copyright system was established to protect works created through the intellectual and creative capacity of the human mind, with originality and creativity forming its fundamental principles. However, with the increasing capacity of artificial intelligence to generate content in areas such as text, visual art, music, or software, the boundaries of this classical system have come under scrutiny.
Due to technological developments, the question of whether a work may benefit from copyright protection now depends not only on the nature of the work itself, but also on whether there was a human contribution in the production process and the extent of such contribution. This is because artificial intelligence may sometimes generate content entirely without human intervention, while in other cases it may be integrated into the production process through human involvement. It is clear that this distinction has a decisive effect on the existence and scope of copyright protection.
Especially in recent years, the approaches and criteria developed by different legal systems on how to make this distinction have attracted attention. In the face of the rapid advancement of artificial intelligence technologies, the question of the extent to which copyright law can adapt to these new forms of creation, and where the boundaries of the system should be drawn, has become increasingly important not only for lawyers but also for all sectors of the creative industries.
2. Artificial Intelligence in Turkish Law
Pursuant to the Law on Intellectual and Artistic Works, a “work” is defined as “any kind of intellectual and artistic product bearing the characteristics of its author, and categorized as literary, musical, artistic, or cinematographic works,” while the “author” is defined as “the person who creates the work.” Although the Law does not explicitly use the term “natural person” in reference to the author, the prevailing scholarly opinion holds that authorship results from creative intellectual labor; therefore, only natural persons may be considered as authors. Entities lacking legal capacity, creative will, and personality—such as artificial intelligence systems—cannot be recognized as authors under Turkish law.
For legal entities to be deemed authors, the creation of the work must occur within the framework of an employment relationship, service contract, or special agreement, and under the instruction and supervision of the legal entity. Even in such cases, the author is still the natural person who creates the work; the legal entity may only acquire certain rights through the transfer of economic rights or by virtue of being the employer.
There is currently no statutory regulation or case law in Turkish law addressing whether artificial intelligence can qualify as an author. In the doctrine, however, the basic condition for protecting content generated through artificial intelligence is that the production must be carried out under human guidance and contribution. Works may be recognized as protected where artificial intelligence is employed merely as a “tool,” meaning that the human directs the process, makes selections, and exercises judgment during production. This approach particularly emphasizes the creative choices made by users in so-called “prompt engineering ” where the user’s contribution determines the final form of the work. In such cases, while artificial intelligence functions as a purely technical assistant, the rights of the person making the creative decisions may be protected. Conversely, in instances where artificial intelligence operates entirely autonomously, without human intervention in the creative process, the resulting content cannot be considered a work subject to authorship.
At this stage, given the absence of specific statutory provisions or judicial precedents in Turkish law, each case must be evaluated individually depending on the nature of the production process. In the near future, technological developments in this field will necessitate legislative and judicial responses.
3. Examples from International Practice
Although Turkish law does not yet provide explicit regulation on copyright protection for content generated by artificial intelligence, developments in foreign legal systems reveal diverging approaches.
In the European Union, the Court of Justice in the Infopaq decision held that, for a work to be eligible for copyright protection, it must constitute “the author’s own intellectual creation.” Accordingly, only human-generated content containing creative choices may be protected, thereby excluding, at least for now, content produced solely by artificial intelligence.
Furthermore, while the 2024 European Union Artificial Intelligence Act does not directly confer copyright protection, it introduces obligations on AI providers to ensure transparency regarding whether models have been trained with copyrighted data, to disclose information to the public, and to grant rightsholders the possibility of objection. These provisions constitute an indirect mechanism of protection, signaling a new attempt to balance innovation with the prevention of copyright infringements.
In China, a similar approach has been adopted. In recent notable cases, the presence of human contribution has been explicitly required for AI-generated content to qualify for copyright protection. For instance, in Li v. Liu, the claimant provided detailed instructions to guide the AI in generating a female portrait. The court held that such detailed guidance amounted to a creative human contribution, recognized the portrait as a protectable work, and declared the claimant Lin as the rights holder.
By contrast, in Feng v. Dongshan, the claimant asserted authorship over a series of AI-generated images. The court, however, concluded that the claimant failed to demonstrate any substantive creative intervention in the production process. The judgment emphasized that merely providing text prompts or selecting images from the AI output would not suffice; for copyright protection to apply, the work must embody genuine elements of human creativity.
Similarly, in the Half-Heart decision, the court stressed that human intervention in the creation process was indispensable for copyright protection of an AI-generated music cover image, reaffirming once again that AI output alone is insufficient for protection.
In the United States, the approach that copyright protection is granted only to works of human authorship is clearly reflected in both legislation and case law.
In the official guidance published by the U.S. Copyright Office in February 2023, it was expressly stated that a copyright registration will not be accepted where the traditional elements of authorship in a work are generated entirely by artificial intelligence. Human contribution limited merely to the idea stage or confined to providing directive prompts during the production process is also deemed insufficient. For a work to be protectable, human intervention must form a concrete part of the creative process
For example, in Thaler v. Perlmutter (2023), the applicant sought copyright registration for an image created entirely autonomously by an artificial intelligence system. The U.S. Copyright Office rejected the application, and this decision was upheld by the federal court. The court emphasized that the term “author” under the Copyright Act refers exclusively to a human being and that this interpretation is grounded in long-standing judicial precedent. The ruling explicitly held that a product created by a machine, without human guidance, cannot be protected.
Another example that came into focus in 2023 was the comic book Zarya of the Dawn, authored by Kristina Kashtanova. In this application, the textual content and the overall arrangement of the comic were created by Kashtanova, while the images were generated through the artificial intelligence system Midjourney. The U.S. Copyright Office recognized the text and the composition created by the human author as protectable, but ruled that the images generated solely via Midjourney lacked sufficient human contribution to qualify for copyright protection. Although it was argued that the prompts contained elements of creative direction, the Office held that such a contribution did not rise to the level of creative authorship and therefore could not give rise to a protectable work.
4. Conclusion and Evaluation
The question of whether content generated by artificial intelligence may benefit from copyright protection is assessed on the basis of similar principles across different legal systems, with the presence of human creativity being recognized as a common criterion for protectability. However, there remain significant uncertainties and regulatory gaps as to how this criterion should be concretized in practice, particularly in cases where human contribution is limited.
The prevailing approach excludes content created in a fully automated manner from copyright protection. Nevertheless, given the increasing role of artificial intelligence in creative processes, it seems inevitable that protection models for such content will need to be reconsidered, and, if necessary, new legal instruments developed. In this context, it is important to establish more flexible and transparent assessment criteria that take into account not only the existence of creative contribution but also its quality and scope.
For copyright law, the challenge will be to strike a new balance—one that neither entirely excludes the presence of artificial intelligence in the creative domain nor abandons the human-centered logic of protection. This will be among the most fundamental legal issues in the near future. For Türkiye, as well, it is essential not to fall behind in these discussions, but rather to adopt predictable regulations aligned with technological developments, thereby filling the gaps encountered in practice.
1 Prompt: A guiding text command or input given to artificial intelligence for the purpose of generating content.
Selin Su, Managing Lawyer
Ebrar Turan, Lawyer
Selen Kaya, Trainee Lawyer