Google’s AI Overviews: A Digital Dilemma Under EU Scrutiny
The European Commission is delving deep into how Google’s AI-driven summaries are reshaping the landscape of digital content under the Digital Markets Act (DMA) and EU copyright laws. This investigation emerges amidst ongoing tensions between AI innovations and traditional internet services, setting the stage for a significant clash.
Revolutionizing Search or Undermining Publishers?
Since the rollout in late March, Google has been responding to search queries with **AI Overviews**—summaries of relevant information that minimize the need for users to visit original content creators’ sites. Google asserts that these summaries will enable users to “find information faster and easier.” However, this shift is causing considerable concern among **news publishers** as they face potential declines in readership and revenue.
As digital businesses live or die by their visibility on Google, this AI-generated approach threatens to distort established business models for publishers, raising questions of competition and copyright that are now under legal scrutiny from the EU.
“The Commission is currently assessing how the ‘Google AI Overviews’ feature works in practice and its implications under EU copyright rules,” stated Commission spokesperson Thomas Regnier in a conversation with Euractiv.
Assessing AI Overviews: A Closer Look at the Findings
While there have been no formal steps initiated under the Digital Services Act (DSA) or copyright directives thus far, the Commission is closely evaluating the impact of AI Overviews as part of its ongoing investigation into Google’s practices under the DMA. Preliminary findings have already indicated that Google may be in violation of the DMA due to self-preferential treatment— favoring its own offerings in search results.
Interestingly, these findings did not specifically address AI Overviews, which were launched after the publication of the initial investigation results. Nonetheless, the underlying principles remain pertinent, as emphasized by Regnier.
Google’s Response
When approached for comments, Google referred to its blog posts discussing its perspectives on AI and copyright, while noting its compliance with both the DSA and the DMA, as well as the EU copyright directive. How this compliance translates into tangible outcomes for publishers remains a pivotal question.
The Web Traffic Tug-of-War
Traditionally, there’s been a reciprocal relationship between publishers and search engines: publishers provide content, and in return, they gain traffic. However, the emergence of AI Overviews appears to tip the balance significantly in favor of the platform.
As Google aggregates more content, many publishers are witnessing a dramatic **decline** in traffic. Bloomberg reported that Google acknowledged this detrimental effect during a private meeting with 20 content creators in October. The BBC shares similar sentiments, suggesting that AI intermediaries might threaten their brand integrity, as they may fail to convey the true value of their reporting or inaccurately summarize their articles.
In a defense of the changes, a senior Google executive pointed to an increase in overall search queries following the introduction of AI Overviews, suggesting a potential boon for publishers. However, when pressed about the implications for **click-through rates**, the executive was not forthcoming with data, indicating that such figures were in a state of flux.
Strategic Timing or Coincidence?
Curiously, the launch of AI Overviews occurred shortly after Google conducted an experiment that showed the removal of European news content from search queries had “no measurable impact” on its revenue. This sync in timing has sparked anger among news outlets, which argue that Google benefits from the credibility of their content despite the low direct revenue from news.
Google claimed that the AI Overviews feature was delayed for about nine months in Europe due to legal uncertainties. It positioned this as a matter of competitiveness for Europe, stating that the pace of innovation is critical and warning that European users could start lagging behind their global counterparts.
Google has argued that EU regulations lead to poorer quality products, with over 20 modifications made to Search to comply with the DMA, as noted by Oliver Bethell, head of competition, regulatory engagement, and advisory at Google.
The Risks of Opting Out
According to **EU copyright directives**, copyrighted material can be utilized for research under a text and data mining exception. Regnier noted that AI-generated summaries could fall under this clause, provided the content is publicly accessible and not under any reservation of rights by the content owner.
However, there’s a caveat: rights holders cannot opt out of being included in AI summaries, only from AI training, creating concern within the industry. Google reassures that opting out of AI training won’t affect a site’s visibility in search results, yet worries linger that this may lead to diminishing visibility for some media outlets.
Renate Schroeder, director of the European Federation of Journalists, emphasized the precarious position many news organizations find themselves in: “We have no choice. We need to be online and visible.”
Schroeder advocates for clearer, more granular disclosures in the AI Act’s mandated transparency, aiming to facilitate rights holders in advocating for compensation for the use of their content. Meanwhile, Google advocates for leniency in transparency requirements, positioning itself at the forefront of discussions surrounding EU AI regulations.
As the Commission’s AI Office mulls over the necessary disclosures for training data, the landscape surrounding Google’s AI Overviews and their repercussions on content creation remains a crucial battleground to watch.
(aw, jp)
