In a never-ending saga between Google and France’s competition authority over copyright protections for news snippets, the Autorité de la Concurrence announced a
€250 millionfine against the tech giant Wednesday (around$270 millionat today’s exchange rate).
The competition watchdog found that Google had disregarded its previous commitments with news publishers, earning the hefty fine. However, the decision took on a new level of significance as it also revealed Google’s use of news publishers’ data for training its generative AI models, Bard and Gemini.
The Autorité de la Concurrence has found that Google failed to notify news publishers of the use of their copyrighted content for training its AI models. This is a breach of previous commitments made by Google to ensure fair payment negotiations with news publishers for reusing their content.
Copyright and competition wrongs
In 2019, the European Union passed a digital copyright reform that extended copyright protections to news headlines and snippets. Google had previously displayed these news stories on their products without providing any financial compensation to news aggregators, such as Google News, Discover, and the “Top Stories” feature box on search result pages.
Google attempted to evade the law by turning off Google News in France. However, the competition authority intervened, stating that this unilateral action was an abuse of Google’s dominant market position and could cause harm to publishers. This forced Google to enter into negotiations with local publishers for content reuse. In 2021, Google was fined $592 million after the competition authority found serious breaches in its negotiations with local publishers and agencies.
The tech giant initially argued that the fine was disproportionate and decided to appeal. However, it later sought to settle the dispute by offering a series of pledges and withdrawing their appeal. These commitments, which were accepted by the French Autorité, include providing key information to publishers and negotiating in a fair manner.
Google has now signed copyright agreements with hundreds of publishers in France – all of which fall under the authority of the agreement with the Autorité. This means that its business in this area is heavily regulated.
While it agreed not to contest the findings of the Autorité in exchange for a fast-tracked process and a monetary payment, it’s clear that Google is still not happy about the situation. In a lengthy blog post, Sulina Connal, Google’s managing director for news and publishing partnerships, wrote that “the fine is not proportionate to issues raised by the authority.” However, Google has decided to settle in order to move on and focus on their goal of finding sustainable approaches to connecting people with quality content and working constructively with French publishers.
GenAI training in the frame
Today’s enforcement by France’s competition authority highlights Google’s use of news publishers’ content to train its AI foundation model, Bard, and its related AI chatbot service, Gemini.
The competition authority found that Google used content from publishers and press agencies to train Bard, its generative AI tool. This tool was launched in July 2023 but was used without notifying the copyright holders or the competition authority. This use of protected content for training purposes is a clear violation of Google’s commitment to the authority.
Google’s defense is two-fold. In their blog post, they stated that the competition authority “does not challenge the way web content is used to improve newer products like generative AI, which is already addressed in Article 4 of the EUCD.” However, the Autorité argues that it has not yet been determined whether this exception applies to Google’s use of news content for training purposes.
The relevant clause in Article 4 of the Copyright Directive specifically states that “reproductions and extractions of lawfully accessible works and other subject matter” can be used for “text and data mining.” It’s worth noting that Google is legally bound to notify copyright holders about uses of their protected works and failed to do so in this case.
Google also briefly mentioned the EU AI Act, suggesting its relevance in this situation. However, this legislation is not yet in force and is still pending final adoption by the European Council.
The incoming AI legislation states that developers must comply with the bloc’s copyright rules. It also introduces transparency requirements, requiring them to have a policy to respect EU copyright law and make a “sufficiently detailed summary” of the content used to train general-purpose AI models, like Gemini and Bard, available to the public.
This requirement may make it easier for news publishers whose protected content has been used for GenAI training to obtain fair remuneration under EU copyright law in the future.
No technical opt-out
The Autorité points out that Google did not provide a technical solution for publishers and press agencies to opt out of their content being used to train Bard without affecting its display on other Google services.
Until September 28, 2023, publishers could use a “noindex” tag to ensure that their content was not used for training Google’s AI model. However, this tag removed their website from Google altogether. In September 2023, Google introduced a “Google-Extended” rule that allowed publishers to opt-out of their content being used for training Gemini’s models specifically, without affecting their display on other Google services. This new rule allowed web publishers to indicate that they did not want to help improve Gemini’s models.
Other shortcomings
The Autorité also sanctioning Google for a number of other issues related to how they negotiate with French news publishers. These include failing to provide adequate information for fair bargaining of remuneration for their content, and using an opaque methodology to calculate payment amounts.
The competition authority also found that Google did not meet non-discrimination criteria, resulting in unequal treatment of publishers. They also criticized Google’s decision to impose a “minimum threshold” for remuneration, stating that this introduced discrimination between publishers. Additionally, Google’s calculations regarding indirect income were not in accordance with previous decisions and the appeal judgment of the Court of Justice from October 2020.
Furthermore, the Autorité noted that Google failed to update remuneration contracts in line with their commitments. In addition, it found that Google’s opt-out processes were not effective and promised to closely monitor them in the future.
