- within Technology topic(s)
- in United States
- with readers working within the Business & Consumer Services industries
- within Technology, Intellectual Property and Environment topic(s)
1. Introduction
The study titled Copyright of AI-generated works: Approaches in the EU and beyond1 (the "Study"), published last December by the European Parliamentary Research Service ("EPRS"), was prepared with the aim of examining how approaches to copyright law in the EU and around the world are evolving in response to the rapidly increasing capacity of artificial intelligence to generate visual, auditory, and written content.
The growing capacity of generative artificial intelligence systems to produce content is pushing the boundaries of traditional copyright law. In conventional copyright regimes, human creativity constitutes the fundamental basis for the protection of a work; however, in the case of AI-generated content, human contribution may be removed from the equation, rendering the concept of "authorship" increasingly ambiguous. The Study examines in particular the approaches adopted by different legal systems in response to the legal uncertainty faced by authors and technology developers alike.
2. Generative AI and Copyright
As discussed in our article dated 18 September 20252, generative artificial intelligence (Generative AI) refers to systems capable of producing new content through the use of large datasets and algorithms. Although these systems generally operate on prompts provided by humans, the resulting "work" is, to a significant extent, generated by artificial intelligence as the outcome of automated processes. In this context, whether such output qualifies for copyright protection is of particular importance, especially as it may constitute a commercial opportunity for artists, designers, engineers, and marketing professionals.
The Study indicates that authors have adopted two different approaches regarding whether works generated by artificial intelligence should benefit from copyright protection. One group of authors argues that works produced solely by artificial intelligence, without any human intervention, should not qualify for copyright protection, as automated processes may diminish the economic value of creative human labour. On the other hand, another group maintains that, in light of the increasing prevalence of AI tools, the use of artificial intelligence in creative processes may foster artistic productivity and technological innovation.
It is further stated that, from a copyright law perspective, the extent to which the use of artificial intelligence is permissible and how the boundaries of human contribution should be determined remain largely uncertain.
3. EU Approach
At the EU level, there is no specific legislation regulating the copyright status of works generated by artificial intelligence. The most recent legislation related to the subject, the Artificial Intelligence Act3 dated 13 June 2024, merely requires AI providers to develop AI systems that comply with EU copyright law.
The gap in the legislation is being filled through the case law of the Court of Justice of the European Union ("CJEU"). In line with the traditional human-centric approach to copyright predominantly adopted by the Member States, the CJEU's jurisprudence generally limits copyright protection to works that reflect the author's free and creative choices. Likewise, this approach is reflected in various European Parliament directives, which state that copyright protection may only arise where there is a meaningful human contribution to the creation of the work.
Ultimately, the EU's current approach is that works generated by artificial intelligence cannot benefit from copyright protection and, accordingly, artificial intelligence itself cannot, as a rule, hold copyright. However, the question of the extent to which human-created works that involve originality and intellectual creativity—albeit with the assistance of artificial intelligence— may benefit from copyright protection remains unclear, according to the Study.
4. Approaches Outside the EU
The Study comparatively examines how legal systems outside the EU approach works generated by artificial intelligence. While some jurisdictions, in parallel with the EU, place human creativity at the core of copyright protection, others have been found to depart from this traditional perspective.
a. The United States of America
In the United States, copyright protection has long been firmly tied to human creativity, and content generated solely by artificial intelligence is not eligible for copyright protection. Indeed, in its January 2025 report titled Copyright and Artificial Intelligence4, the U.S. Copyright Office stated that outputs generated by AI systems based on user prompts are, on their own, insufficient to render the prompt provider a copyright holder. However, it was also noted that those aspects of the output that contain a sufficient degree of human creativity may benefit from copyright protection. The Study emphasizes that the question of which specific elements of AI-generated outputs may qualify for protection has not yet been fully clarified5.
b. Common Law Countries
In the United Kingdom, New Zealand, and Hong Kong, the concept of "computer-generated works" is accepted, and copyright in such works may be attributed to the person who regulates and enables the creation process.
c. China
The approach in China links copyright ownership to works created by natural persons or works created under the control and responsibility of a legal entity. The study observes that Chinese courts have achieved a first by interpreting copyright rules in light of AI outputs; however, no consensus has yet been reached regarding the level of human intervention required in the creation of the work.
d. Ukraine
In Ukraine, copyright legislation provides a sui generis form of protection for images generated by artificial intelligence, distinct from the traditional copyright approach, while maintaining the classical copyright regime for the parts of a work created by a human. Accordingly, copyright protection is granted to the authors of the computer program underlying the AI system, their heirs, the persons to whom the economic rights in the program have been assigned, or the lawful users of the program. The Study notes that a sui generis copyright regime similar to that adopted in Ukraine has been rejected in most EU Member States due to concerns that it could undermine competition in the intellectual, cultural, and creative sectors.
5. Comparative Analysis
The Study reveals significant differences in approach among various legal systems. Nevertheless, the prevailing view places human creativity at the core of copyright protection and refrains from directly incorporating AI-generated outputs into the traditional copyright framework on the grounds that they lack "human authorship". An example of this situation is given in the terms of use of artificial intelligence systems such as Gemini and ChatGPT6, where the copyright ownership of images created as artificial intelligence outputs is not accepted.
On the other hand, the Study notes that the rapid development and increasing number of AI systems may necessitate different approaches in light of global competition. As the level of AI involvement becomes progressively more difficult to detect, greater transparency and documentation at various stages of the creative process may become essential, potentially prompting legislators to introduce alternative forms of protection to prevent copyright infringements.
6. Evaluation
The main conclusion reached in the Study is that there is no clear and harmonized global approach regarding the copyright status of works generated by artificial intelligence. Although there is no explicit regulation at the EU level, the existing case law and the policy of the European Commission support a human-centred understanding of copyright.
Different approaches at the international level create legal uncertainty, particularly in cases of cross-border use, and point to the need for more concrete regulations that will protect the rights
of authors in a balanced manner in the long term. Indeed, the "Suryast" case clearly demonstrates the extent to which different approaches have been adopted in India7, Canada8, and the United States9 with regard to copyright legislation for a single output produced with the help of artificial intelligence.
In conclusion, the Study emphasizes that the copyright status of works generated by artificial intelligence is not merely a theoretical debate; rather, it has direct legal and economic implications for individuals operating in the creative industries, developers of AI systems, users, and regulatory authorities.
Although Turkish doctrine and the existing legal framework, similar to the approach adopted in the EU, closely associate copyright protection with human creativity, the increasing prominence of generative artificial intelligence systems in creative processes demonstrates that this approach may prove inadequate in the long term. In this context, it is considered that initiating a discussion under Turkish law on a sui generis protection model, outside the scope of the classical copyright regime, aimed at safeguarding economic interests in AI-generated outputs may become inevitable.
Footnotes
1 European Parliamentary Research Service, Copyright of AI-generated works: Approaches in the EU and beyond, EPRS Document PE 782.585 (December 2025)
2 Selin Su, Selen Kaya ve Ebrar Turan, 'Yapay Zekâ ve Telif Hakkı', Moral (18 September 2025)
3 Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 on a framework for artificial intelligence and amending certain Union legislative acts (AI Act) (OJ L 1689, 13 June 2024)
4 U.S. Copyright Office, Copyright and Artificial Intelligence Part 2: Copyrightability (January 29, 2025)
5 United States Copyright Office, Letter Regarding Zarya of the Dawn (Registration # VAu001480196) (21 February 2023)
6 OpenAI, Terms of Use (1 January 2026)
7 Sarkar, S. (2021, August 5). India recognises AI as co-author of copyrighted artwork. Managing Intellectual Property.
8 Samuelson-Glushko Canadian Internet Policy and Public Interest Clinic v Ankit Sahni, (8 July 2024), Ottawa, FCC T-1717-24
9 Copyright Review Board, Second Request for Reconsideration for Refusal to Register SURYAST, Copyright Review Board, December 11 2023
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
