A Munich regional court ruled that OpenAI's ChatGPT violated German copyright law by training its language models on protected song lyrics from popular artists without permission, siding with GEMA (Germany's music rights society representing approximately 100,000 creators). The court rejected OpenAI's argument that language models "absorb training sets" rather than copy specific works, and ordered the company to pay undisclosed damages. The decision is the first major European precedent establishing that AI tool operators must comply with copyright law during training, and GEMA signaled intent to negotiate compensation frameworks with OpenAI. For financial institutions, the ruling creates quantifiable training-data IP liability for AI vendors and establishes precedent that non-compliance incurs both damages and remediation costs in major markets.
271 other tracked developments touching the same claims.