A Munich court has dealt a major blow to OpenAI, ruling that the artificial intelligence giant infringed on copyright law by using protected song lyrics to train its ChatGPT models without permission. Tuesday’s verdict represents a landmark moment in AI regulation in Europe and may change the way tech companies license content across the continent.
The court ruled that OpenAI infringed on artists’ rights by both storing song lyrics in its language models and reproducing them in the outputs of the chatbot. This dual finding is particularly important because it covers the entire AI pipeline, from training to output generation.
The legal challenge came from GEMA, Germany’s influential music rights organization representing more than 100,000 composers, songwriters, and publishers. Filed in November 2024, the lawsuit focused on nine German songs whose lyrics are allegedly used without proper licensing or compensation for their creators.
OpenAI Loses Copyright Case in Munich Over AI Training and Output
GEMA argued that OpenAI had no right to reproduce these protected works for commercial purposes without first obtaining licenses and paying the artists who created them. The organization stressed that creators are entitled to reasonable remuneration if their work is being commercially exploited and that this enables them to make a living.
The San Francisco-based company’s defense was based on two main arguments. First, OpenAI argued that its language models do not actually store or copy specific data; instead, the company contended, such models simply reflect learned patterns in their settings, a technical distinction that the court ultimately rejected.
Second, OpenAI tries to diffuse responsibility by stating that users who produce outputs through ChatGPT are the originators, thus should be held liable legally; this argument was also denied by the court.
Munich judges were crystal clear in their judgment: “Both the memorization in the language models and the reproduction of song lyrics in the chatbot outputs are copyright infringements.” They concluded that the plaintiffs were entitled to compensation both for the reproduction within the models themselves and in the generated outputs.
OpenAI reacted by expressing disagreement with the decision and stated that it was “considering next steps.” The company downplayed the ruling’s significance, noting that it concerns “a limited set of lyrics” and does not affect the millions of Germans who use their technology each day. OpenAI stressed its respect for creators’ rights and ongoing conversations with organizations worldwide.
However, legal experts and industry groups view the impact of the ruling rather differently. Raue, the law firm representing GEMA, called the verdict “an important precedent for the protection of creative works” that sends “a clear signal to the global tech industry.”
German Copyright Ruling Delivers “Milestone Victory” Against AI Companies
GEMA noted that this was the first big copyright case of its type in Europe against an AI company, while OpenAI faces various lawsuits from media organizations and authors in the United States.
Kai Welp, GEMA’s head of legal affairs, said he hoped the ruling would prompt AI companies to negotiate fair compensation agreements. “Authors need remuneration for the commercial exploitation of their works so that they can make a living,” Welp said.
The impact of the ruling is larger than music. GEMA said the ruling could extend into other forms of creative content, meaning that AI companies might have to approach everything from written articles to visual art differently.
The German Journalists’ Association welcomed the decision enthusiastically, calling it “a milestone victory for copyright law.” This reaction suggests that journalists and other content creators are watching closely, potentially emboldened to pursue similar legal challenges.
The Raue law firm noted that the ruling provides “legal certainty for creative artists, music publishers and platforms throughout Europe and is likely to have an impact far beyond Germany.” This prediction seems well-founded, as other European countries and organizations may now feel more confident pursuing copyright claims against AI companies.
With artificial intelligence continuing to evolve at breakneck speed, this ruling by the German court represents a momentous point in the setting of boundaries with regard to how those AI companies use copyrighted material.
The balance between technological innovation and intellectual property protection remains a highly contested territory, but what this ruling signals is that AI companies cannot just help themselves to creative works without proper licensing and compensation.
