Authorship in the Age of Artificial Intelligence: A Civil Law Approach
Résumé
The emergence of automated processes in the creation of artistic works has become an undeniable reality. Thanks to applications that fall under what mainstream is calling “generative AI”, it is now possible to generate images from mere descriptions or to construct texts based on concise instructions (prompts). As a result, the issue of authorship of these creations has regained prominence in legal discourse. The present study therefore seeks to shed light on this very question.
First, we will disregard the machine-centric perspective and instead adopt a human-centric approach. Accordingly, our analysis will focus on the automated execution process to determine whether a work can exhibit originality, i.e., the manifestation of free and creative choices made by a human author.
Second, we will analyze the notion of "imprint of data" with the goal of establishing that the replication of certain patterns from pre-existing data does not prevent an individual from using artificial intelligence systems to create original works.
Finally, we will turn our attention to determining the author of an original AI-generated work. In doing so, we will identify the primary candidates as the user and the developer, and then evaluate the circumstances under which each party might be considered appropriate for such attribution.