Google AI Overviews Dispute Is Now a Formal EU Antitrust Investigation
The dispute over Google’s AI-generated search answers is no longer limited to a publisher complaint. On December 9, 2025, the European Commission opened a formal antitrust investigation into Google’s use of online content for AI Overviews and AI Mode, examining content terms, compensation and access for competing AI developers, as set out in its official investigation notice.
As of August 13, 2026, the proceeding remains an investigation rather than a finding that Google infringed EU competition law. Google has since started a limited test that lets some website owners exclude their material from generative Search features, but the experiment does not settle the inquiry’s broader questions about payment, earlier practices or the position of rival AI companies.
What publishers challenged in 2025
The original dispute concerned the connection between visibility in conventional Google Search and participation in its generated answers. Reuters’ July 2025 report on the filing described a complaint dated June 30, 2025, from the Independent Publishers Alliance, the Movement for an Open Web and Foxglove Legal, together with a request for interim measures.
The complainants alleged that Google placed summaries produced with publisher material above conventional links, reducing the prominence of the original pages. Their more fundamental objection was about control: they argued that a publisher could not refuse the relevant crawling or use while retaining normal eligibility to appear in Google’s general search results.
That distinction matters because declining participation carried a potentially wider cost than losing exposure in AI Overviews alone. The complaint framed this arrangement as an abuse of Google’s power in online search and warned of serious, irreversible harm to independent media. Google disputed the account of publisher traffic, arguing that Search sends websites billions of clicks and that fluctuations can also reflect changing demand, user interests and algorithm updates.
The formal inquiry reaches beyond referral traffic
The later EU investigation overlaps with the publishers’ complaint but covers a wider competitive question. It examines whether web content was used for AI Overviews and AI Mode without appropriate compensation or an effective opportunity to refuse, and whether Google imposed unfair conditions or gave itself privileged access that placed competing AI developers at a disadvantage.
A separate branch concerns YouTube. Regulators are examining the use of videos and other uploaded material to train Google’s generative AI models where creators may lack both compensation and a meaningful opt-out, while YouTube’s rules restrict competing model developers from obtaining comparable access.
The inquiry therefore is not simply an attempt to measure how many visits a generated answer may displace. It concerns the terms on which a dominant search provider acquires content, uses that material in its own AI services and controls access to an important source of training data.
Opening an investigation is not a judgment against Google. It establishes the practices and legal questions under examination, but it does not prove an abuse of dominance or guarantee any particular remedy. Possible outcomes could differ from the interim restrictions and publisher controls requested by the original complainants.
Google’s opt-out test changes one part of the dispute
Google’s June 3, 2026 website-owner update says a new Search Console toggle is being tested with a subset of UK sites, allowing them to exclude links and content from AI Overviews, AI Mode and AI Overviews in Discover. Sites that opt out receive neither impressions nor visits from those generative features, while the choice is not used as a ranking signal for results outside them; the same limited rollout also introduced reporting on impressions, pages appearing in AI responses and the countries where they appeared.
This test addresses a central part of the control problem by separating participation in generative results from ranking elsewhere in Search. It still requires publishers to surrender any visibility and potential traffic available inside the AI features themselves, so it is not a way to withhold content while preserving every existing source of discovery.
The control is also narrower than the regulatory inquiry. It does not determine whether publishers should be compensated when their work contributes to generated answers, whether the conditions applied before the test were fair or whether competing AI developers had access on equivalent terms. Its limited UK scope also means it should not be described as a generally available global opt-out.
What has changed—and what remains unresolved
Publishers now face two related but distinct developments. The antitrust investigation concerns whether Google’s conduct complied with EU competition rules; the Search Console experiment concerns the product controls Google is prepared to provide voluntarily. A broader rollout could alter publishers’ operational choices without deciding the legal case, while an eventual regulatory decision could demand measures beyond the current test.
The June 2025 complaint did not immediately produce a ban on AI Overviews, mandatory licensing payments or a universal participation control. It was followed by a broader formal investigation, and Google later began testing a separate opt-out for generative Search with a limited group of UK website owners.
The unresolved issue is therefore more specific than a general conflict between artificial intelligence and journalism. Regulators must determine whether a powerful search service can use third-party material in generated answers under terms that satisfy EU competition law, including the treatment of publishers, creators and rival AI developers.
Also read:
Subscribe to our newsletter
Get the latest Web3, AI, and crypto news delivered straight to your inbox.