ChatGPT defeat: GEMA wins lawsuit against OpenAI over song lyrics usage
Munich. The Munich Regional Court has put the brakes on OpenAI and its artificial intelligence ChatGPT. The US corporation's use of song lyrics violates copyright law, according to the judges.
They thus upheld a lawsuit filed by the German collecting society GEMA, which had specifically sued over nine songs – including well-known titles such as "Atemlos" and "Männer" by Herbert Grönemeyer, "Über den Wolken" by Reinhard Mey, and "In der Weihnachtsbäckerei" by Rolf Zuckowski. The ruling is not yet legally binding.
OpenAI is alleged to have used song lyrics without a license.The texts had been used to train ChatGPT and were returned exactly, or at least very similarly, in response to simple queries to the system. The court considered this proof that the texts had been stored in OpenAI's systems.
Among other things, it ordered OpenAI to refrain from storing the texts and outputting them in its models, to pay damages, and to disclose information about the use and the revenue generated from it.
It is considered likely that the ruling will be appealed and will involve further appeals. The final decision could have implications far beyond song lyrics, as expert Silke von Lewinski from the Max Planck Institute for Innovation and Competition explained before the ruling.
She sees "fundamental importance for all works, be it literature, journalistic texts, music, visual arts, photography or any other works used for generative AI. This is about how to interpret the laws that already exist."
She was convinced that if GEMA were to win in the final instance, this would shift the balance of power between the creative industries and technology companies somewhat in favor of authors and other rights holders.
"Before a text can be used for generative AI, the rights holders would have to give their consent and would have the opportunity to receive compensation for it."
This is likely also the aim of GEMA with its lawsuit.
That the AI was trained with the nine songs was undisputed in the trial (case number 42 O 14139/24). What happened afterward, however, was a key question. Was the song data memorized—that is, saved and thus duplicated—or did the training with the data cause ChatGPT to regenerate the lyrics without saving them?
The court took a clear stance, interpreting the fact that the system output the texts it had been trained on as proof that it must have memorized them. Random output was ruled out.
RND/dpa
rnd
