Artificial Intelligence and Copyright: Challenges and Opportunities
AI technologies and their relationship with Intellectual Property (IP) rights have remained a prominent topic on the global agenda this year, as they were last year. With the increasing presence of generative AI models such as ChatGPT, BERT, LaMDA, and DALL-E2, which are capable of creating original content by processing data like text, images and sound, debates continue over whether the outputs generated by these models qualify as copyrighted works, who owns the rights to them if they do, and potential copyright infringements. In this context, while numerous disputes have been brought before courts worldwide, countries have also initiated legislative efforts to address the legal gaps in this area.
Indeed, the AI Act[1], prepared by the European Union as the first known legal regulation in this field, entered into force on August 01, 2024, and will become fully applicable on August 02, 2026. Although this Act does not directly address copyrights, it can be said that the Act in question is a regulation that provides specific, transparent mechanisms for the use of AI technologies, respecting the rights of third parties - including copyright rights- and deterring violations by imposing sanctions.
Moreover, the UK Government published a consultation on December 17, 2024 on changes and clarifications to copyright law to resolve the debate on AI and it will run until February 25, 2025.[2] The consultation outlines the main objective of identifying a balanced approach that both protects rights holders and human creativity, while fostering innovation in the AI sector and ensuring legal certainty for long-term growth. Among the issues discussed are proposals such as the disclosure of work and data used in AI training, transparency regarding web browser details including ownership and purpose, content licensing and the requirement to maintain records to demonstrate compliance with rights reservations. Additionally, the consultation questions whether existing legal frameworks adequately align with the evolving technology and contemporary needs, particularly in providing copyright protection for computer-generated works.
In addition to the legislative work, disputes and lawsuits regarding AI and copyrights continued to arise in many countries this year.
A group of artists has filed a class action against Stability AI, Midjourney, DeviantArt and Runway AI, claiming that their works were used without permission as training data for AI and/or provided to users as outputs, resulting in violations of their rights, along with allegations of encouraging infringement and unfair enrichment through the system provided to users. Although no decision has been rendered yet on the merits of the dispute, the judge of the case dismissed the other claims of the plaintiffs in the preliminary examination, but decided to continue the proceedings with respect to the allegations of copyright infringement by unauthorized use of the plaintiffs' works as AI training material, obtained from the internet, in respect of the AI model named Stable Diffusion.[3]
The NewYork Times also sued OpenAI and Microsoft for copyright infringement, claiming that its content was provided to users through ChatGPT and Copilot AI search results without authorization.[4] The case discusses whether creating AI models without permission with copyrighted works can be considered fair use. The plaintiff's allegations that the defendant OpenAI has deleted the evidence[5] that forms the basis of the lawsuit from its own systems raises concerns about the specific challenges these AI systems may pose in judicial processes, collection of evidence and examination of the subject matter of the dispute due to their technical structures.
The attitude that the courts adopt in these cases will be crucial for the development of the related field and will set a precedent for similar cases.
Another decision by the Guangzhou Internet Court in China has also attracted considerable attention. In this case, the court ruled that certain images generated by the AI model based on users' text inputs were substantially similar to the famous Japanese cartoon character “Ultraman”. The court found that the AI provider had accessed the original works through an online database and infringed upon the exclusive licensee's reproduction and adaptation rights through the images created as a result of the relevant inputs. The court also ordered the cessation of the infringement and the payment of damages.[6]
While global developments in disputes and legislative efforts are ongoing, many discussions in this field continue within our country. Although we have yet to see any legislative work or court decisions on AI and copyright disputes, it is clear that such developments will become inevitable in the coming years, given the widespread use of these technologies in daily life and the emerging needs they bring.