OpenAI has been ruled by the Munich Regional Court to have violated German copyright laws by reproducing song lyrics from popular artists. The lawsuit was brought by GEMA, a German rights organization representing over 100,000 creators, which claimed that OpenAI's language models memorized and reproduced lyrics from nine specific songs when prompted. The court found that the lyrics were embedded in OpenAI's models, affirming that the responsibility lies with the company, not the users generating the prompts. OpenAI has pushed back, arguing that its models do not store or copy specific training data. They maintain that outputs are a result of user input. Despite this, the court emphasized that the defendants duplicated the lyrics and made them publicly accessible, which infringes on copyright exploitation rights. GEMA’s CEO, Tobias Holzmueller, criticized the notion that creative work on the internet is free to use without compensation. The ruling is part of a broader discussion around copyright in the age of generative AI. Similar cases are ongoing, including GEMA’s lawsuit against Suno AI and previous legal battles involving other AI companies like Stability AI and Anthropic. These cases highlight the growing tension between AI development and intellectual property rights, as companies navigate the complexities of using copyrighted materials for training their models.