As artificial intelligence continues to reshape the global economy, tensions between technology developers and content creators are intensifying, particularly around how training data is sourced and compensated. Arthur Mensch, CEO and co founder of Mistral, has proposed a new framework for Europe that would require AI model providers to contribute financially to the cultural sector through a revenue based levy.
The proposal comes amid growing legal and regulatory scrutiny of how AI systems are trained on large volumes of publicly available data, including text, audio, and video content. Creators and copyright holders have increasingly raised concerns about the use of their work without compensation, while AI companies face uncertainty over compliance with evolving regulations.
A proposed levy for AI providers
Mensch’s proposal centres on the introduction of a levy applied to companies that develop or deploy AI models within Europe. The contribution would be based on revenue generated from AI services and would apply to both European and international providers operating in the region.
The aim is to create a consistent framework that ensures all companies benefiting from access to publicly available content contribute to the ecosystem that produces it. By applying the levy equally to domestic and foreign firms, the model seeks to establish a level playing field within the European market.
Funds collected through this mechanism would be directed into a central European fund dedicated to supporting content creation and strengthening cultural industries.
Supporting creators and innovation
The proposed fund would be used to invest in new content production and provide financial support to creative sectors across Europe. This approach is intended to address concerns that creators are not adequately compensated for the use of their work in AI training processes.
At the same time, the framework aims to preserve opportunities for direct licensing agreements between creators and AI companies. Rather than replacing existing commercial arrangements, the levy would function as a complementary system, ensuring baseline contributions while allowing more specific partnerships to continue.
By linking AI development with support for cultural production, the proposal attempts to balance technological progress with the sustainability of creative industries.
Addressing regulatory fragmentation
One of the key motivations behind the proposal is the current regulatory landscape in Europe, which Mensch describes as fragmented and uncertain. While existing rules allow AI companies to use copyrighted material for text and data mining under certain conditions, the lack of clarity has created challenges for both developers and rights holders.
In contrast, AI companies in other regions are often operating under more permissive or less defined frameworks. This difference has raised concerns about the competitiveness of European AI companies, which may face additional legal and operational complexities.
The proposed levy is positioned as a way to provide greater legal certainty, reducing the risk of disputes while establishing clear expectations for all participants in the market.
Legal certainty for AI developers
A central element of the proposal is the idea that companies contributing to the levy would benefit from protection against certain liabilities related to the use of publicly available content for training purposes. By participating in the system, AI developers would gain a clearer legal framework within which to operate.
This could help accelerate innovation by reducing uncertainty and enabling companies to focus on developing new technologies without the constant risk of legal challenges. At the same time, it offers a structured approach to addressing the concerns of content creators.
A broader debate on AI and copyright
Mensch’s proposal reflects a wider debate taking place across Europe and beyond regarding the relationship between AI and copyright. As AI systems become more capable and widely adopted, questions about data usage, ownership, and compensation are becoming increasingly urgent.
European policymakers are already exploring long term solutions to address these issues, with ongoing discussions about how to balance innovation with the protection of intellectual property.
The concept of a revenue based levy represents one potential approach, aiming to align the interests of AI developers and content creators while supporting the broader ecosystem.
Looking ahead
While the proposal has yet to be formally adopted or confirmed in detail, it highlights the growing need for new frameworks that reflect the realities of AI driven economies. As the technology continues to evolve, finding sustainable models for data usage and compensation will be critical.
By advocating for a system that combines financial contributions, legal clarity, and continued commercial flexibility, Mensch is contributing to a broader conversation about how Europe can shape the future of AI in a way that supports both innovation and cultural production.
