A top EU court adviser has warned that Google's Gemini chatbot may violate copyright when it reproduces press content in search answers. This non-binding opinion signals a potential shift where tech giants might need to pay licensing fees for AI content, posing a long-term regulatory risk for AI-driven business models.
Google is facing a new regulatory hurdle in Europe that could impact how its AI models like Gemini operate. An adviser to the Court of Justice of the European Union, Advocate General Maciej Szpunar, has shared a legal opinion suggesting that while AI systems can legally use public data for training, the way they present information to users might cross a legal line.
The opinion draws a sharp line between two AI functions. It notes that using copyrighted works to build or train AI models is generally allowed under existing EU rules. However, the situation changes when an AI chatbot produces answers that mimic or repeat protected content from publishers. According to the adviser, if a tool essentially mirrors a unique article or helps a user get around a paywall to read paid news, it could be seen as an illegal reproduction of work.
This distinction is vital for investors because it challenges the current path of AI search development. If the full court adopts this view in its final ruling, companies like Google may face a shift in how they design and manage their services. It could force tech firms to negotiate licensing agreements with media houses or update their tools to prevent the reproduction of specific proprietary content. Either path could increase operating costs or require significant changes to the way Gemini summarizes and presents information to users.
While this opinion is currently non-binding, the European court often follows the advice of its Advocates General. The final verdict, expected in about six months, will likely set a key legal precedent for how generative AI is regulated across Europe. Investors should monitor the progress of this case, known as Like Company v. Google Ireland, as the outcome may dictate how AI search products are monetized and whether they can continue to rely on open access to digital news content.