Will Europe’s First AI Music Trial Force Tech to Pay Up?
It could pit GEMA against OpenAI in a battle that decides whether AI companies must license music, setting the tone for how creative work is used across Europe.
A Munich court is about to hear a first-of-its-kind case that could decide how artificial intelligence (AI) and music coexist in Europe. At the center is German rights group GEMA, which has taken OpenAI to court, accusing the company of copying lyrics it represents without permission.
At first glance, you might think this is just another dispute over song lyrics. But if you look closer, it’s the first major test of whether AI companies in Europe will need licenses to use copyrighted music for training. And unlike in the U.S., where “fair use” leaves plenty of gray space, European law is often far stricter. That’s why this ruling could set a precedent with industry-wide consequences.
GEMA’s complaint is pretty direct. It says its songwriters and publishers had opted out of AI training, yet OpenAI allegedly scraped lyrics from online sources anyway. In some cases, ChatGPT even generated songs word-for-word. To GEMA, that’s not inspiration but reproduction. OpenAI is expected to counter that not every source of lyrics is covered, that outputs aren’t the same as copies, and that training models shouldn’t require a license in the first place.
Emmanuel Umahi
And this fight isn’t happening in isolation. Earlier this year, record labels sued AI startup Suno for allegedly pirating songs from YouTube to train its models. In Latin America, musicians have warned that bots are hijacking streams with machine-made tracks, eating into royalties and threatening cultural lifeblood. If you’ve been following these stories, you can see how quickly this battle is escalating.
That’s what makes the Munich case so important. If GEMA wins, we could see billion-euro licensing deals between tech firms, labels, publishers, and collecting societies, essentially creating a new market for AI training data, similar to how Spotify pays artists for streams. If OpenAI prevails, developers can keep training without deals, at least for now. But that would only deepen tensions with the music industry, and it’s hard to imagine regulators in Brussels staying on the sidelines.
The implications also stretch far beyond music. The same logic could be applied to novels, scripts, photography, or visual art. If licensing becomes the norm, AI won’t just change how companies build models; it could also reshape how creative industries monetize their archives.
Europe is clearly moving faster than the U.S. on these questions, and this case could be the first domino to fall. As GEMA’s general counsel Kai Welp put it: “We want to put pressure on the companies to acquire licenses.” Whether or not the judges agree, that pressure is already building, and it’s only going to grow.
Emmanuel Umahi
