The right to compensation and commercial activities: is a balance possible in the development of artificial intelligence?

Abstract

With the rapid development of artificial intelligence technologies, the question of the legality of using authors’ works in the process of creating artificial intelligence is increasingly being raised. In this article, the authors seek to strike a balance between the interests of commercial artificial intelligence developers and the right of authors to fair remuneration. The study assessed the legality of the artificial intelligence training process, taking into account the legal norms and court practices of various jurisdictions. It has been established that in most countries, training artificial intelligence using copyrighted works is a legitimate practice. Different countries apply different regulatory models – some provide broader opportunities, while others impose stricter restrictions. However, there is a clear imbalance between the desire of artificial intelligence developers to freely use works and the right of authors to fair remuneration. This article seeks to find a way to reconcile these opposing positions. Based on the concept of “what is worth copying is prima facie worth protecting”, the authors propose an alternative interpretation – “what is worth copying is prima facie worth remunerating” – and emphasise the need for a compensation mechanism.
Existing and theoretical compensation models proposed by scientists that could be applied in the future are also analysed. After evaluating each of them, it was determined that a mechanism that is administratively effective and financially sustainable should be adopted. The authors present what they consider to be the most effective model, based on several analysed and discussed models. However, the conclusion is that the most likely model to prevail is the one based on a one-off compensation license. This is determined by the opt-out option for text and data mining established in the European Union and the so-called Brussels effect, which influences the direction of artificial intelligence regulation on a global scale.