In the wake of ongoing work on training data transparency and remuneration models, the Conseil supérieur de la propriété littéraire et artistique (CSPLA) has entrusted Tristan Azzi, professor at the Université Paris 1 Panthéon-Sorbonne, and Yves El Hage, lecturer at the Université Jean Moulin Lyon 3, with a mission as strategic as it is complex: to determine the applicable law, in terms of private international law, for the training of generative AI models commercialized within the European Union.
These two experts were given this mission on April 23, the same day the government announced the launch of a consultation cycle between representatives of generative AI model developers and those of rights holders in the cultural and media sectors, led by Marc Bourreau and Maxime Boutron. Although distinct, these two initiatives pursue a common goal: to ensure the protection of the rights of creators and neighboring rights holders.
Coming into force in August 2024, the Artificial Intelligence Regulation (AI Act) establishes a legal framework for the development, marketing, and use of AI systems, in line with European Union values. It specifically requires providers of general-purpose AI models destined for the European market, even if offered for free, to respect intellectual property rights, even if their model's training was conducted outside the EU. They are thus required to identify and comply with the rights reservations expressed by holders of protected works, in accordance with Directive (EU) 2019/790 on copyright in the digital single market, known as the DSM Directive.
The aim of this extraterritorial approach is to prevent AI providers operating in countries with less protective copyright or neighboring rights laws from gaining an unfair advantage, to the detriment of creators and rights holders.
The Objectives of the Mission
Assisted by Yves El Hage as rapporteur, Tristan Azzi has been tasked with:
- examining the state of regulation and case law applicable at both national and international levels;
- presenting the conflict of law issues posed by AI development in terms of copyright and neighboring rights;
- studying the various options to clarify the handling of these conflict of laws to safeguard the interests of authors and neighboring rights holders.
Their conclusions are expected in December 2025 during the final plenary session of the year of the CSPLA.