In recent years, artificial intelligence (AI) has developed by leaps and bounds, having started from a theoretical notion contemplated by Alan Turing in 1950 about the ability of machines to think like humans, followed by the development of basic programmers in strategy games and natural language processing; then, in recent years where the greatest evolution has been seen, such as GPT-3 from OPEN AI, where AI demonstrates advanced capabilities for generating natural text, and in 2023 with Chat GPT based on GPT-4, capable of generating coherent and natural text, and finally in 2024, where AI with generative capacity can be observed, becoming an omnipresent technology that impacts and transforms various sectors such as medicine, education, entertainment, manufacturing processes, and that which we will focus on in this article: INTELLECTUAL PROPERTY.
In this regard, intellectual property refers to the creations of human beings, such as inventions, artistic works, literary works, musical works, works of art, trademarks, and others, which must meet the requirement of originality to be protected under copyright, thus becoming high-value assets for their author and for companies as they represent a competitive advantage over third parties.
It is for this reason that AI and intellectual property are linked, since, with the training of AI bases and its generative content capability, several questions arise regarding intellectual property, including: Does AI infringe copyright by having bases trained with protected works? Can the results (outputs) obtained from the instructions (prompts) given be subject to copyright?
These questions have been raised and do not have an absolute answer, since, as mentioned at the beginning, AI has evolved by leaps and bounds, and intellectual property laws have lagged behind, currently existing a legal vacuum in this regard in countries such as Colombia.
However, in other parts of the world, such as the European Union, regulation on this matter has begun, leading globally in how this issue will be addressed legally, thus establishing the requirements and obligations regarding the uses of AI, trustworthy use, unacceptable risks, prohibited AI practices, and the safety and protection of the fundamental rights of individuals and companies.
With that regulation, the European Union's stance on the aforementioned questions can be seen: first, regarding copyright, the regulation allows the training of AI from protected content without authorization, provided that the rights holders do not reserve their rights, in which case AI model providers must obtain authorization to feed their bases with the texts and data of the works; also, as a measure, it is established that AI providers must prepare and make available to the public a detailed summary of the content used for AI training, being careful regarding the protection of secrets and confidential business information, so that authors can exercise and enforce their rights.
Similarly, regarding the question of whether the results will be subject to copyright, the AI Act does not contemplate a solution, but the European Parliament has discussed the issue, analyzing the feasibility of conferring legal personality on AI, and in that sense granting it the copyright of the resulting works, this being the option with the least likelihood, or the option that certain requirements be met so that the results can be protected, such as the holder being a human being, originality, and respect for pre-existing intellectual property rights, so that in this case, where creations are human and AI has been assisted, intellectual property rights are conferred on the human being, as would not happen with works produced autonomously by AI, since this is not linked to a human creative process.
In the same vein as the European Union, in the United States, regarding the first question, safeguarding creativity and copyright in the age of AI is contemplated through transparency regarding the information used to feed AI before it reaches the market, so that authors can speak out, give their consent or not, and thus their rights are not transgressed.
Next, regarding the question of whether copyright is generated in the results obtained by AI, it is evident that the stance of United States judges is that it is not, as in the case initiated by Stephen Thaler, owner of the software called “Creativity Machine,” who generated an image through this artificial intelligence and requested registration with the Copyright Office, where the request was rejected, which gave rise to a lawsuit bringing the case before the Federal Court of Washington D.C., which ruled rejecting the registration, arguing that art generated by this type of technology cannot be protected under copyright, as it lacks human authorship.
The aforementioned argument has been reiterated to deny the registration of other works, such as in the case of the comic Zarya of the Dawn, where the Copyright Office, in addition to the above, adds that although the registration applicant Kristina Kashtanova was the one who gave instructions to the AI, these do not function as orders but as suggestions, without there being control over the finally generated image; similarly, the registration of the photograph, a 2-D artwork, was denied, where the tool RAGHAV Artificial Intelligence Painting App (“RAGHAV”) was indicated as the author of the 2-D artwork.
Thus, it can be observed how some countries are establishing their stance on copyright and AI, without setting aside the continuous analysis to determine the best option and regulation in this regard, being a topic open to discussion, since cases of registration requests and copyright infringement continue to arise, and more questions emerge, such as: Are AI results original? Can the instructions (prompts) given to AI be subject to intellectual protection? Can works resulting from instructions given to AI, with modifications, be protected? Is transparency regarding the information used for AI training and the possibility for authors to reserve their rights a sufficient measure? And others that generate uncertainty on the matter.
Therefore, and as an alternative, despite it being a new topic that is just beginning to be regulated, the creators of these platforms have chosen to provide some security to their users, as in the case of the COPILOT platform, owned by Microsoft, within which, in its terms and conditions, Microsoft commits to its users to respond in case of intellectual property claims and to assume responsibility for the associated legal risks.
In short, the growing influence of Artificial Intelligence in our society confronts us with a series of unprecedented opportunities and challenges; as this field advances by leaps and bounds, it is crucial to recognize the need for updated regulation that addresses the complex ethical, legal, and social implications that this entails, as leading countries such as the European Union and the United States have begun to do, taking significant steps towards legal harmonization, which in turn serves as a guide for other countries, such as Colombia, towards a future where AI is used ethically and responsibly for the benefit of humanity.