Mistral chief proposes EU levy on AI firms to fund creative industries

Mistral is calling on European governments to tax AI companies 1-1.5% of revenues to compensate creators whose work trained the models. Firms paying the levy would gain legal protection from copyright liability claims.

Categorized in: AI News Creatives
Published on: Mar 21, 2026
Mistral chief proposes EU levy on AI firms to fund creative industries

Mistral Proposes European Levy on AI Companies to Compensate Creatives

French AI developer Mistral called Friday for European governments to impose a revenue-based levy on companies selling AI models, with proceeds funding cultural industries and content creators.

Mistral's chief proposed a tax between 1.0 and 1.5 percent of revenues from AI systems trained on human-generated content - text, audio, and video pulled from the public web. The company said the money would flow into a central European fund for new content creation and cultural sector support.

The proposal addresses a core complaint from creators and copyright holders: AI developers have trained systems on vast amounts of copyrighted work without direct compensation or permission. Legal challenges have mounted in both the US and Europe, with authors and publishers arguing their work fueled AI development.

Mistral framed the levy as a path to legal certainty. AI companies paying the tax would gain protection from liability claims over training data sourced from publicly available online materials - without eliminating direct licensing agreements between data owners and AI firms.

The company emphasized the levy would apply equally to foreign AI developers operating in Europe. Most major AI companies are US-based and operate under what Mistral called "extremely permissive regulatory contexts on copyright."

The EU's AI regulation, adopted in 2024, requires systems to respect copyright rules. But how those rules apply to generative AI systems remains unsettled, leaving creators in a gray zone.

The Copyright Problem

Mistral itself faced scrutiny last month when French media reported the company used copyrighted works including "Harry Potter" and "The Little Prince" to train its models. Mistral said it respects opt-out mechanisms and deploys safeguards against copyrighted material, though some works are "especially popular and duplicated many times online," making complete exclusion difficult.

In September, AI giant Anthropic agreed to pay $1.5 billion to settle a class-action lawsuit by authors - one of the few concrete compensation arrangements reached so far.

Mistral's proposal sits between two positions: it acknowledges creators deserve compensation while offering AI companies a fixed, predictable cost structure instead of ongoing litigation. Whether European regulators adopt such a mechanism remains unclear.

For AI for Creatives, understanding how compensation models might work is increasingly relevant as Generative AI and LLM systems become standard industry tools.


Get Daily AI News

Your membership also unlocks:

700+ AI Courses
700+ Certifications
Personalized AI Learning Plan
6500+ AI Tools (no Ads)
Daily AI News by job industry (no Ads)