The Munich court has ruled in favor of GEMA against OpenAI, finding that ChatGPT had reproduced protected song lyrics without authorization. This is a historic turning point, which nevertheless does not resolve all the issues. Yesterday, the Munich Regional Court handed down a long-awaited judgment in the case pitting German collecting society GEMA against OpenAI. The 42ᵉ civil chamber, which specializes in copyright law, ruled in favor of GEMA on most of its claims, finding that OpenAI’s language templates had reproduced protected song lyrics without authorization. The ruling, which is not final and which OpenAI has announced it will appeal, marks a first European interpretation of the scope of the so-called « text and data mining » (TDM) exception – introduced by the 790/2019 copyright directive – to the field of training artificial intelligence systems. TDM: exploitation without reproduction As a reminder, Article 4 of Directive 790/2019 states that « Member States shall provide for an exception or limitation to rights (…) for lawfully accessible reproductions and extractions of works and other protected subject matter for the purpose of text and data mining. » In its statement, the Münich court explains that « when, as…
