In a never-ending saga between Google and the French competition authority over copyright protection for news snippets, the Autorité de la Concurrence announced On Wednesday, the tech giant was fined 250 million euros (around $270 million at today's exchange rate).
According to the competition watchdog, Google has breached a few of its previous commitments to news publishers. But the choice is especially notable since it ignores something else entirely current – taking a cue from Google's use of reports publishers' content to coach its Bard/Gemini generative AI model.
The competition authority has accused Google of failing to tell news publishers about GenAI's use of their copyrighted content. This comes against the backdrop of previous commitments from Google geared toward conducting fair payment negotiations with publishers for the reuse of their content.
Copyright and competition violations
In 2019, the European Union passed an EU-wide digital copyright reform that prolonged copyright protection to headlines and clippings. News aggregators corresponding to Google News, Discover, and the Top Stories feature box on search results pages had previously scraped these news stories and displayed them on their products without financial compensation.
Google originally wanted to bypass the law by shutting down Google News in France. But the competition authority quickly intervened, deeming its unilateral motion an abuse of market dominance that would harm publishers. The intervention essentially forced Google to enter into content reuse deals with local publishers. But in 2021, Google was fined $592 million after the competition regulator found serious violations in its negotiations with local publishers and agencies.
The technology giant called the sanction “disproportionate” and said it could appeal. However, it subsequently attempted to resolve the dispute by making quite a lot of undertakings and withdrawing its appeal. The commitments were accepted by the French Autorité and include the transfer of necessary information to publishers and fair negotiations.
Google has signed copyright agreements with tons of of publishers in France that fall throughout the scope of its agreement with the Autorité. Therefore, business on this area may be very strictly regulated.
No appeal
Google has agreed to not challenge the Autorité's latest findings in exchange for an expedited trial and a monetary payment.
However, managing director of reports and publishing partnerships Sulina Connal struck an indignant tone – she wrote in a single long blog post that “the positive is disproportionate to the problems raised by the authority”.
The blog post suggests that Google really desires to draw a line under the story this time, with Connal also writing: “We have reached an agreement because it’s time to move on, and as our quite a few agreements with publishers show, we would like to “Let’s give attention to the larger things.” Goal of sustainable approaches to connecting individuals with high-quality content and constructive collaboration with French publishers.”
With generative AI on the horizon and competition to adopt tools, Google's calculus for solving the content reuse problem looks different.
GenAI training within the context
Today's enforcement by the French Competition Authority shows that it has tightened Google's use of content from news publishers and agencies for training purposes for its AI Foundation model and associated AI chatbot service Bard (now Gemini).
It found that Google used content from publishers and press agencies to coach Bard, its generative AI tool launched in July 2023, “without notifying copyright holders or the agency,” its report said Press release.
Google defends this point in two ways. In her blog post, she writes that the competition authority “doesn’t query the best way web content is used to enhance latest products corresponding to generative AI, which is already regulated in Article 4 of the EUCD (EU Copyright Directive)”.
Article 4 of the Copyright Directive sets out an “exception or limitation for text and data mining” – particularly for “reproductions and extractions of lawfully accessible works and other material for the needs of text and data mining”.
However, the Autorité argues in its press release that it just isn’t yet clear whether the exception applies here. (It's value noting that the relevant clause refers to “lawfully accessible works” – whereas Google has a legal obligation to the competition authority to tell copyright holders in regards to the use of their protected works, and appears to have did not achieve this on this case.)
“As to the query of whether using news content to coach a man-made intelligence service falls throughout the scope of protection, this query has not yet been clarified,” wrote the competition authority. “However, the Autorité considers that Google breached its Commitment No. 1 by not informing publishers that their content was used to coach Bard.”
Google's blog post also mentions the EU AI law in passing – suggesting that it’s relevant. However, the law just isn’t yet in force because it remains to be pending final adoption by the European Council.
The upcoming AI laws may even require developers to comply with the bloc's copyright rules. And with this goal in mind, transparency requirements are being introduced: they have to introduce a directive to comply with EU copyright law. and make publicly available a “sufficiently detailed summary” of the content used to coach general AI models (corresponding to Gemini/Bard).
This latest requirement for model builders to publish a summary of coaching data could make it easier in the longer term for news publishers whose protected content has been included for GenAI training to receive fair compensation under EU copyright law.
No technical opt-out
The Autorité also points out that Google has failed to offer it, no less than until then September 28, 2023a technical solution that enables publishers and press agencies to opt out of using their content for Bard training without such a call affecting the display of their content on other Google services.
“Until this date, publishers and news organizations that desired to opt out of this use case were required to incorporate a directive blocking all Google content indexing, including for its Search, Discover, and Google News services. “These services are specifically a part of the negotiation of revenue related to related rights,” it said, adding: “Going forward, the Autorité will fastidiously consider the effectiveness of Google’s opt-out processes.”
Technically, between July and September 2023, news publishers could add a “noindex” tag to the robots.txt file to be certain that their content just isn’t used to coach Google’s AI model. This robots.txt file is placed in the basis folder of web servers and comprises various instructions for search engines like google and yahoo. Google's web crawler examines the instructions in these files to index web sites.
However, a “noindex” tag means your site will disappear from Google completely. In September 2023, Google added more granularity and created a “Google-Extended” rule, which is different from the “noindex” rule. By opting out of the Google Extended Statement, web publishers indicate that they don’t need to contribute to improving Gemini's current and future models.
Other defects
The Autorité can also be imposing sanctions on Google over quite a lot of other issues related to the best way the corporate negotiates with French news publishers, finding that the corporate didn’t provide them with all the knowledge essential to barter fairly To ensure remuneration for his or her content.
In his Press releaseIt said that the knowledge provided by Google to publishers in regards to the method used to calculate the compensation they might be paid was “particularly opaque.”
Google was also found to have failed to satisfy non-discrimination criteria geared toward ensuring equal treatment of publishers. And she pointed to Google's decision to set a “minimum threshold” for compensation, i.e. Below a certain threshold, all publishers are “arbitrarily allocated zero remuneration, no matter their respective situation,” the press release continues.
In addition, the Autorité criticized Google's calculations of so-called “indirect income” and stated that the “package” it proposed was not in step with previous decisions or the Court of Justice's October 2020 appeal ruling.
It also said Google had not followed through on its commitment to update compensation contracts in step with its commitments.
