There is no consolidated definition of AI. We can say, however, that it is a branch of computer science that aims to create machines and programmes that perform tasks like a human being. This requires large amounts of data, which will be processed by these machines or programs, thus enabling a certain level of automatic learning and autonomous decision-making.
Problems arise with generative AI, such as ChatGPT, Dall-E, Midjourney etc. In accordance with the request of the user, these programmes are able to generate texts or images based on collected information from a wide variety of sources. They do not take into account, however, any legal elements, such as possible intellectual property rights of third parties, or even privacy and data protection.
In this regard, the Austrian authors’ association Autorinnen Autoren has already called for AI programs and systems that can write texts, such as ChatGPT, to be blocked temporarily. On the grounds that such programs can copy literary styles and plagiarise writings, they argue for the block not to be lifted before their use is compatible and in line with existing copyright regulations.
This is not the first, nor the last case of this type: the number of lawsuits against AI companies for violating intellectual property rights when training their systems is noticeably growing. Possible solutions to this problem could be to establish algorithms in AI programs that take into account the copyrights of the materials and data they process to generate content, or even to pay royalties to those authors.
Looking at the issue from a legal point of view, we must call on the Spanish Law on Intellectual Property (Ley de Propiedad Intelectual). This law aims to protect works created by artists, authors or other creators, and to ensure that they are not used without the prior authorisation of the creator. Furthermore, the law establishes that creations are protected by copyright if they are original. Such originality requires the creator to be a human being.
Consequently, “works” created by generative AI do not meet the legally established requirements to be protected by copyright, as they cannot be considered works and do not meet the prerequisite of originality. Works shall be considered original, according to the decision of the CJEU in the case C-145/10, they are a “intellectual creation of the author reflecting his personality and expressing his free and creative choices in its production.”
Therefore, if such AI generated “creations” are not protected by intellectual property, they might be considered to be in the public domain. If this were the case, anyone could use them for their own interests. In spite of this, there are voices advocating for some kind of protection of these types of creations.
The reason for this advocacy might be that these new types of AI-driven creativity have very significant commercial implications. If they are not protected by copyright, and can be used freely, investments in these types of technologies are likely to decline eventually because they are not profitable. There would no point in investing millions in technology to produce certain creations, if afterwards there was no monopoly on their use and commercialisation. We shall see.
Unaffected by these challenges is the copyright protection of computer-generated works, because the programme is merely a tool to support the creative process. However, in the latest types of AI, the programme is no longer a mere tool, it makes decisions without human intervention.
Clearly, many questions regarding this issue still require an answer. In light of this legal vacuum, the European Union is preparing a draft law aimed at regulating the use of AI (expected in 2024) to resolve the current uncertainty. But this is only the beginning. As the use of AI becomes more widespread and machines produce better creative works, the situation will become even more complex not only for the rights of authors, but even for the safety of citizens.
This technology will be able to clone images, videos and voices of real people, with a high level of detail. With this technology, innocent people could be charged with criminal acts that have not occurred but have been generated by AI, with the violation of fundamental rights that this would entail.
In any case, we will be following the developments closely to see how this technology evolves and whether states finally begin to regulate its development and application.