by Emilio Berkenwald

General interestJanuary 8, 20260 comments 0 Likes

Towards a new balance between human creation and AI

Artificial intelligence has become a central part of our daily lives. It generates texts, images, melodies, and even contributes to the development of technical innovations. However, this advance has sparked a legal and philosophical debate: can AI be the owner of intellectual property rights?

For now, the short answer is no. The world’s leading intellectual property offices require that the inventor or author be a human being. In the case known as “DABUS,” patent applications were filed naming as the inventor an AI (“Device for the Autonomous Bootstrapping of Unified Sentience”) developed by researcher Stephen Thaler. Between 2023 and 2025, courts in the United Kingdom, Germany, Korea, and Japan confirmed that only humans can be recognized as inventors, even if the system generated the invention autonomously. The case became a global benchmark and confirmed that, at least for now, artificial intelligence cannot hold industrial property rights, particularly invention patents.

Some authors argue that, in the near future, it may be necessary to redefine the notion of legal personality to include machines with autonomy and creative capacity. They even speak of a possible “robotic legal personality” that would allow certain rights or obligations to be recognized for intelligent systems. For now, that idea seems closer to science fiction than legal reality.

So, what can be protected by intellectual property rights in the context of artificial intelligence? It is important to distinguish between the different types of protection that exist. If AI generates an artistic or literary work, the result could be protected by copyright, depending on the territory, provided that there is a certain degree of human intervention in the creative process. In the case of an industrial design, its ornamental aspect may be protected, and in the case of a technical invention, a patent could be applied for. In all cases, the author of the work, the designer, or the inventor must be human beings, and the objects must meet the corresponding validity requirements (originality in the case of works; novelty and distinctive character for a design; novelty, inventive step, and industrial application for an invention).

In the case of patents, current practice recognizes “AI-assisted inventions,” i.e., developments in which artificial intelligence collaborates with humans and is used as a tool, but does not replace their role as inventors. Therefore, patents for AI-assisted inventions continue to be examined under the same criteria as computer-implemented inventions, for which there is extensive case law. Under this approach, it is also possible to protect generative AI technology involving a model or algorithm, provided that it meets the traditional requirements of novelty, inventive step, and industrial application.

In turn, since patents are not authorizations for use, there may be obstacles to exploitation when the development obtained with AI is based on training data or software protected by prior rights. A researcher may hold a patent on a new AI application, but if they used data or tools protected by third parties to develop it, they may need to obtain the appropriate authorizations in order to exploit it commercially.

This tension is evident in the debate over Studio Ghibli-style images. In 2024, various AI models began producing visual works that imitated the Japanese studio’s style, prompting complaints from creators’ associations and warnings about the unauthorized use of protected material in model training. In some jurisdictions, a “visual style” is not always protected by copyright per se, although the production of derivative or overly similar works could constitute infringement, e.g., if they reproduce well-known characters or scenes.

Some countries are already considering whether to establish specific exceptions for the use of data for training generative AI models. The idea would be to allow certain uses without infringing copyright, provided that conditions such as transparency, scientific purpose, or non-commercial use are met.

Artificial intelligence poses a challenge to the current legal framework. Its ability to create, learn, and evolve forces us to rethink the boundaries of authorship and intellectual property. We are in the process of redefining the relationship between humans and technology in a context where machines are no longer just tools, but creative actors with a real impact on the economy and culture. At the same time, the law will have the difficult task of finding a balance between protecting innovation, ethics, and common sense.