This article aims to scrutinize two interconnected issues related to the use of GenAI tools in the film industry, namely the copyright protection of outputs and inputs, from both a regulatory and case law perspective, and comparing them in the United States, Europe, and Italy. It will also address the impact of private regulations, the so-called Terms of Use (ToU), on these issues. Recent regulations and practices of the US Copyright Office concerning the most problematic scenarios (including generic and detailed prompts, expressive inputs, and the editing, selection, and arrangement of AI-generated content) will be taken into consideration in arguing for the reasonableness of solutions and their potential application across the Atlantic in light of European rules and case law. The analysis of the ToU reveals not only previously undetected risks but, even worse, a strong ‘Washington Effect’. This conflicts with European AI regulation and copyright law, particularly regarding moral rights and the right to opt-out in AI training, and the possible infringement of copyrighted works used as expressive inputs. Finally, the article compares the conclusions of the theoretical investigation with the results of a survey conducted among film industry professionals. The empirical analysis reinforces the validity of the theoretical study, confirming concerns raised by the analysis of the ToU, where principles of fairness and the consequences of the Washington Effect are at stake.
Francesca Vessia (Thu,) studied this question.