EU court weighs if Gemini's news summaries breach copyright law

The legal challenge in Europe

Google is currently facing a legal challenge in Europe regarding how its Gemini chatbot interacts with news content. The case was brought forward by Like Company, a website operator from Hungary, to the Court of Justice of the European Union (CJEU). This dispute highlights the growing friction between technology platforms and media organizations over the use of copyrighted journalism.

The core of the issue involves how generative AI tools use news articles without what many publishers consider to be adequate permission or payment. As chatbots become more common, media companies are worried about the implications of their content being used to create automated responses. This legal battle seeks to define the boundaries of fair use in the age of artificial intelligence.

The outcome of this case is being watched closely by the global tech industry. A ruling against Google could have significant financial consequences for AI platforms. It might also open the door for a wave of compensation claims from various publishers whose content fuels AI-generated answers.

How the copyright rules work

Advocate General Maciej Szpunar provided an opinion on the case, suggesting that publishers' rights could be infringed under specific circumstances. He noted that a violation could occur if a chatbot retrieves an article from a website and reuses it to generate a response. This is particularly relevant if the chatbot circumvents paywalls or other technical restrictions to access the content.

The court must distinguish between the facts reported in a news story and the specific way those facts are expressed. While copyright protection does not extend to the raw facts themselves, it does protect the unique expression of those facts. If an AI response reproduces recognizable elements of a publication's original expression, it may be considered a breach of copyright rules.

The opinion also makes an important distinction between training an AI model and generating a response. The use of copyrighted material to train large language models (LLMs) is generally permitted under the European Union's text and data mining exception. However, the court is looking at what happens when the AI actually presents a response to a user based on that content.

Google's defense and position

Google has responded to these concerns by emphasizing the importance of European law's text and data mining exceptions. The company argues that these rules are vital for the competitiveness and development of AI within the region. They maintain that their technology is designed to learn linguistic patterns rather than simply replicating the content of publishers.

The company asserts that Gemini produces original responses based on the patterns it has learned during its development process. They argue that the tool is not intended to copy the specific work of journalists but to provide information to users. This distinction is central to their argument that their technology does not violate the rights of content creators.

Furthermore, Google points out that publisher controls are already in place to give rights holders the ability to manage their content. These tools allow publishers to decide how their information is used or to opt out of certain processes entirely. Google suggests that these existing mechanisms provide a way for publishers to maintain authority over their own intellectual property.

Impact on the news industry

Media organizations are currently facing growing pressure as AI platforms become more integrated into how people find information. There is a widespread concern that AI-generated summaries are significantly reducing traffic to original news websites. When a chatbot provides a direct answer, users may no longer feel the need to click through to the source to read the full story.

The European Publishers Council has voiced concerns about this shift in online discovery. They argue that there is a need to distinguish between search engines indexing news and AI-generated answers that substitute for original reporting. The council believes that publishers should not have to choose between being visible on search platforms and protecting their copyright.

Angela Mills Wade, the executive director of the council, highlighted Google's significant influence over how audiences discover content online. She argued that the company's role in directing traffic makes the issue particularly significant as AI-generated answers expand. The council seeks a balance where publishers can maintain their visibility without losing control over their copyrighted work.

What to watch for next

The Court of Justice of the European Union is expected to deliver its final ruling on this case in the coming months. While the opinion provided by Advocate General Maciej Szpunar is non-binding, the court typically follows the majority of such opinions. The final decision will likely set a major precedent for how AI companies interact with news content across the continent.

A ruling in favor of the publishers could lead to new licensing models where AI companies must pay for the right to summarize news. This would change the economic landscape for both technology companies and the media organizations that provide the underlying data. It could also force AI developers to implement more rigorous filters to ensure they are not reproducing recognizable elements of copyrighted work.

Conversely, a ruling that favors the current AI practices would provide more certainty for the expansion of generative AI tools. It would reinforce the idea that training and generating responses based on public information falls under existing exceptions. Both outcomes will have a lasting impact on how information is consumed and distributed in the digital age.