歐盟法院顧問指 Google Gemini 或侵犯新聞版權

An adviser to the EU’s highest court has put forward an opinion suggesting that Google’s Gemini chatbot could violate copyright laws within the bloc. Gemini is an artificial intelligence model designed to generate responses to user queries. The potential infringement, as outlined, specifically relates to the chatbot's practice of incorporating press articles into the answers it provides.

The core of the adviser's opinion, according to Reuters, lies in a critical distinction: the legality of training AI models on copyrighted material is separate from the act of reproducing publishers’ content in the actual responses given to users. This implies that while the initial learning phase of an AI model might be treated differently, the output it generates is under scrutiny for direct copyright violations.

For companies developing and deploying AI chatbots like Gemini, this opinion suggests a potential need to scrutinize how their models synthesize and present information sourced from copyrighted works. If a chatbot's answer directly reproduces portions of a publisher's article, it could fall under infringement, even if the model's training on that article was not the primary concern. This shifts the focus of potential liability to the user-facing output.

It is important to note that this is an opinion from an adviser, not a final, binding ruling from the EU’s highest court. However, such an opinion can often influence the court's ultimate decision, setting a precedent or clarifying how existing copyright laws might apply to AI-generated content within the European Union.