The main objective of the research is to analyze the possibility of legal protection of outputs of AI under copyright and related rights law, or sui generis rights, to establish what would be the best option for the regulatory framework of the European Union. Moreover, it will investigate who the rightholder would be, in the case that the output of AI technology can be considered a work under copyright law.
Central to the analysis would be to study the actual function of a supposed intelligent software. If the AI really had some kind of intelligence, could a human being who controls it actually have a creative role? As copyright law has included expressions that go further away from its original scope, like software, how much of an actual creative process is required? And when does this creative process start and finish? Which kind of human participation is there in the creative process in which there is an interaction with AI?
The following methodology will be applied: first of all, there will be an analysis of the technical aspects of AI and neural networks, and their definitions, as there are no fixed criteria within the scientific community. Following, there will be a comparative analysis of the originality and authorship criteria in copyright laws of the continental and Anglo-Saxon systems, and of their proposals to regulate AI outputs not protected by copyright.
Afterwards, the research continues with a breakdown of the basis for protection for copyrightable works and related rights. As copyright is deemed to be an incentive or a reward for the creation of works, how does this apply to the cases of creation with AI? Is there a need to incentivize AI developers giving them copyright over the AI outputs? It should be considered that most of the technical AI researchers are based in the US and China, and those jurisdictions have not implemented any special copyright law in relation to this kind of creation.
Then, a study of eventual reasons to modify the legislative framework from a European perspective will follow, taking into account both legal and technical as well as policy reasons. In particular, this part will study whether there should be a specific sui generis system for the protection of AI outputs at EU level, and whether the EU normative framework should be modified to establish conditions under which outputs of AI cannot be classified as a copyrightable work, in order to maintain legal certainty. Finally, it will analyze whether certain conditions should be established so that a legally acquired work can be used as a source (input) for the processing by AI.