/ Artificial Intelligence and Copyright: Challenges and Opportunities in the Digital Age
March 11, 2025Senior Associate
Alessandri
Artificial intelligence (AI) presents a significant challenge in the field of intellectual property (IP), particularly concerning copyright law. In its current state of development, AI can autonomously generate scientific, artistic, or literary content without human intervention. These materials may exhibit forms of expression considered original and, therefore, could be classified as intellectual works protected by copyright.
AI is making remarkable progress in the technological and business sectors, being implemented across a wide variety of industries. Its impact extends to multiple aspects of the creative process[1]. The growth of AI is driven by the availability of large volumes of training data and technological advancements that allow for high and affordable computational capacity. AI interacts with IP in multiple ways, affecting both the creation and protection of intellectual works.
The issue of AI-generated works primarily concerns the ownership of the rights derived from such intellectual creations. The central question is determining who should be granted the rights over these works. However, the debate first focuses on a more fundamental question: whether an object autonomously created by AI qualifies as an intellectual work that can be protected under copyright law.
Indeed, various jurisdictions have already ruled on cases where copyright protection was sought for AI-generated works, determining that such works are not eligible for copyright protection due to the absence of human intervention in their creation[2]. For instance, in the United States, Judge Beryl A. Howell ruled that works created entirely by an AI system without any human involvement are not eligible for copyright protection. This establishes human intervention as an indispensable requirement for a work to be recognized as a protected intellectual creation.
The discussion on copyright ownership for AI-generated works requires, first and foremost, clarifying whether human intervention is essential for an object to be legally considered an intellectual work. This debate is crucial, as the definitions of authorship and originality in the context of AI pose significant challenges to the current legal framework.
As AI becomes more deeply integrated into the creative field, the distinction between human-created works and machine-generated ones is becoming increasingly blurred[3]. Advances in computing and computational power have enabled AI systems to learn and replicate artistic styles from vast datasets, mimicking human creations with growing accuracy[4].
The ability of AI to generate creative works presents complex challenges in the realm of intellectual property. With sufficient computational power and training data, we may soon struggle to distinguish between content created by humans and that produced by machines. This scenario forces us to reconsider existing legal frameworks and decide what kind of protection, if any, AI-generated works should receive, especially when human intervention is minimal or nonexistent[5].
Currently, copyright law maintains the criterion of originality, and only natural persons are recognized as copyright holders. The key issue is whether there is an economic justification for creating a general copyright exception to foster AI development—a topic under legislative discussion in Chile and worldwide. Additionally, there is debate over whether AI-generated works should be considered derivative of the software owner—who develops and trains the AI system, the user who provides the prompts to generate the content, or the copyright holders of the works used to train the AI—raising questions about whether AI owners should pay licensing fees to human creators. Depending on whether AI system owners are required to license copyrighted training materials, the licensing terms for end users could include either of the following alternatives: the resulting work belongs to the “pro” user who provided the prompt, or to the AI owner, who can define different usage rights (e.g., private, commercial, professional, educational).
A pragmatic approach, supported by various stakeholders, suggests granting copyright to the individuals or entities responsible for enabling AI operations. This model—similar to the approach proposed in the United Kingdom—ensures that companies continue investing in AI technology, with the confidence that they can obtain a return on their investment[6]. This approach would also promote transparency and accountability regarding the use of copyrighted data for AI training.
The future debate will likely center on whether computers should be granted rights similar to those of humans—a discussion that remains to be fully explored.
___________
[1] Article “Inteligencia artificial y propiedad intelectual”, WIPO – https://www.wipo.int/about-ip/es/frontier_technologies/ai_and_ip.html [2] Article “IA y derechos de autor: la sentencia que marca un precedente”, MARCASUR Magazine – https://marcasur.com/noticia/ia-y-derechos-de-autor-la-sentencia-que-marca-un-precedente&f=08-2023 [3] Artificial Intelligence and Copyright, WIPO Magazine, October 1, 2017 – https://www.wipo.int/es/web/wipo-magazine/articles/artificial-intelligence-and-copyright-40141 [4] Article “Humanos Prefieren Arte ‘Real’ Que el de la IA, Pero Les Cuesta Diferenciarlo” – https://decrypt.co/es/210567/humanos-prefieren-arte-real-que-el-de-la-ia-pero-les-cuesta-diferenciarlo [5] Article “La IA frente al toque humano: Un análisis comparativo de la creación artística” – https://www.domestika.org/es/blog/12176-la-ia-frente-al-toque-humano-un-analisis-comparativo-de-la-creacion-artistica