The Justification for Establishing Exceptions and Limitations to Copyright for Programs based on Artificial Intelligence

Abstract

The issue of artificial intelligence (‘AI’) in the context of intellectual property law, including copyright law, has attracted continued interest. Progressive innovation brings new challenges, and the advances we have seen in recent years - particularly in the development of generative artificial intelligence (‘GenAI’) systems – are attracting media and public attention. The adoption and use of generative artificial intelligence systems has sparked widespread debate about their relevance to the copyright system. In the wake of emerging questions, copyright holders have begun to file copyright infringement lawsuits against artificial intelligence companies targeting the process of training artificial intelligence with the results obtained from generative artificial intelligence systems. As a result of these questions, copyright holders have begun filing copyright infringement lawsuits against owners of programs trained on the basis of data protected by copyright and data protection law. Drawing on analysed discussions, normative proposals, consultations and recommendations from experienced practitioners, this article identifies one of the broad questions of contemporary copyright policy towards artificial intelligence, concerning the legality of using copyrighted works to train artificial intelligence models. It also poses the question of the desirability of establishing a new system of copyright exceptions and limitations dedicated to artificial intelligence systems, while analysing the impact of existing limitations under copyright exceptions and limitations on the development of artificial intelligence.