15 EU States Demand New AI Copyright Law for Training Data

Human Verified

Share to: , and

Three European Union flags, a gavel and law books on a desk in front of a parliament building

Germany and fourteen other European Union states have asked the European Commission to propose a new copyright law for the age of artificial intelligence, covering training-data transparency, stronger enforcement of creators’ rights and fair pay for the work AI systems are built on.

What the States Asked For

The call came from Germany’s culture minister Wolfram Weimer at a meeting of EU culture and media ministers in Brussels on 20 November. France, Poland and Spain are among the fourteen states joining Germany, and the group wants the Commission’s legislative proposal as soon as possible.

The states want creators and rights holders paid for the work AI systems feed on, plus transparency about the data used to train models and a level playing field for AI companies. Weimer put it bluntly: creative variety must not be a free raw material for the AI industry.

Why Now

The demand lands as the Commission prepares its own copyright package, scheduled for the second quarter of 2027. The states’ letter calls the Commission’s targeted initiative for a better copyright environment an important step, but says Europe still needs a full regulatory framework behind it.

The same Brussels meeting also covered youth protection online, including the EU Kids Act and age limits for social media, and adopted a new EU work plan for culture running to 2030. Germany also backed a French push to strengthen Europe’s media landscape and an ambitious revision of audiovisual rules.

Our Read

Fifteen states is a heavyweight signal, and France joining matters most. The earlier Council note that circulated this week listed fourteen countries including Germany but not France; this is a different, larger call, and it puts the Commission under real pressure to act.

The ask is political, not a law. The Commission can take the proposal up, water it down or sit on it, and labs’ training-data practices will not change until a text becomes law. But a joint call from fifteen governments is hard for Brussels and the labs to ignore.

The timing fits a wider regulatory turn. US senators introduced a bill this week requiring AI companies to disclose their financing and data practices, and the European push now adds copyright law to a list of rules labs can no longer treat as optional.

Why It Matters

For AI labs, training-data transparency is the fight that matters most. If Europe forces disclosure of what went into models, it could reshape licensing deals, dataset choices and the legal exposure of every frontier system trained on copyrighted work today.

For creators, the call is about pay. Weimer’s line — that creative work must not be a free raw material — frames the debate as a labour question, not a technical one, and that framing is what gives the fifteen states’ letter its political force.

Background

Copyright and AI training has been contested all year. Wikipedia’s parent drew its own line in October over fair use, and courts and regulators on both sides of the Atlantic are still working out where training ends and infringement begins. The fifteen states want a law to settle it.

Source: German Federal Government

Share to: , and

Faheem R. Khattak

Senior automation engineer and AI marketing journo covering LLMs, algorithms, analytics, core updates, Bing search, and technical SEO.

0 Comments

Leave a Reply

Optional Take Tag:

🛡️ Spam protection: Honeypot + Time-Trap Active

This site uses Akismet to reduce spam. Learn how your comment data is processed.