Why two AI copyright cases ended so differently—and what creators need to know now

DCM Editorial Summary: This story has been independently rewritten and summarised for DCM readers to highlight key developments relevant to the region. Original reporting by The Conversation, click this post to read the original article.

Artificial intelligence companies and creative industries are facing off in courts around the world over copyright issues. More than 60 lawsuits are currently active in the U.S., with artists and rightsholders accusing AI firms of using their work without permission. Recent rulings in the UK and Germany have highlighted how courts are interpreting copyright law in this new technological era, and these decisions could shape future developments.

In the UK, Getty Images sued Stability AI, claiming the company used its content to train the Stable Diffusion AI model without permission. The judge acknowledged that Getty’s images were used and copied during the training. However, because the AI training occurred outside the UK, claims were limited to what’s known as “secondary infringement”—similar to how importing counterfeit physical goods is treated under the law. The court had to decide whether digital AI models could count as “articles” under the law, and if they represented infringing copies. In the end, the court decided that the AI model did not store actual copies of Getty’s content, resulting in a loss for Getty.

Germany’s court took a different stance. GEMA, a major music rights organization, sued OpenAI for using song lyrics in training ChatGPT. The court ruled that any AI training involving copyrighted content requires a proper license. Unlike in the UK, the German court focused less on the technical details and more on the end result—that the AI could reproduce the lyrics it was trained on. This led to a win for GEMA.

Despite these different outcomes, both courts agreed on an important point: it’s the AI developers who are legally responsible if copyright is violated, not the users prompting the AI. These early rulings show how different countries are interpreting copyright law in the context of AI. With several other cases still pending and appeals likely, the legal landscape may continue evolving.

Ideally, this conflict won’t be settled only through the courts. A better outcome would involve AI companies and the creative industries working together through proper licensing agreements. Such collaboration should ensure fair compensation and consent from creators, whose work fuels both the creative and tech worlds.