Publishers' Antitrust Claims Against Google's AI Overviews Collapse in Court
A federal ruling exposes the gap between traffic expectations and monopoly law, leaving the generative search question open.
The Core Ruling
A federal court in the United States has thrown out two separate lawsuits accusing Google of using its AI-powered search features to divert traffic from digital publishers and education platforms. Judge Amit Mehta, the same jurist who ruled in 2024 that Google maintained an illegal monopoly in online search, concluded that neither plaintiff demonstrated a violation of antitrust law. The decision underscores a crucial distinction: possessing market dominance does not automatically render every business practice unlawful.
Penske Media Corporation, which operates titles including Rolling Stone, Variety, and Deadline, filed suit alongside education technology firm Chegg. Both argued that Google's AI Overviews feature repackaged their content in search results, answering user queries directly and eliminating the need for click-throughs. The plaintiffs claimed this practice damaged their advertising revenue by reducing site visits while simultaneously using their material as free training data for Google's generative models.
Mehta's memorandum opinion dismantled the legal foundation of both complaints. The plaintiffs, he wrote, had alleged only an expectation that Google would send them search traffic in exchange for making their content available. That expectation, however reasonable it might feel to publishers who have optimised their sites for Google's algorithms over two decades, does not constitute a binding agreement. "An expectation is not an agreement," Mehta observed. "It is simply how a general search engine works."
The Coercion Argument That Wasn't
Chegg's complaint included a second claim: that Google was effectively coercing websites into allowing their content to be scraped for AI training by threatening exclusion from search results altogether. This argument had intuitive appeal. Publishers face a binary choice under Google's current policy: either allow your pages to appear in AI Overviews, or use a technical directive (a specific robots.txt instruction) that permits traditional indexing but blocks generative use. The latter option still allows pages to appear in conventional search results, but as AI features consume more screen real estate on results pages, that visibility becomes less valuable.
Mehta found no antitrust violation here either. The judge concluded that Chegg failed to show Google was leveraging its search monopoly to gain an unfair foothold in a separate market, such as digital publishing or education content. Nor did the plaintiffs establish that Google's conduct amounted to an illegal tie - forcing acceptance of one product (AI scraping) as a condition for another (search indexing).
The ruling reflects a longstanding principle in competition law: dominance in one market does not prohibit a company from entering adjacent markets, even when doing so creates tension with existing players. What the law prohibits is using monopoly power in an anticompetitive manner to foreclose competition or coerce participation.
The Traffic Compact Nobody Signed
At Opentechwire, we have tracked the implicit bargain underpinning the commercial web since Google's PageRank algorithm reshaped online discovery in the early 2000s. Publishers structured their businesses around search engine optimisation, accepted Google's crawlers, and built advertising models predicated on a steady flow of referral traffic. Google, in turn, built the world's most valuable company by indexing that content and displaying snippets alongside advertisements.
That arrangement was never formalised in contracts, yet it functioned as a de facto compact for more than twenty years. The introduction of generative AI search features has destabilised that equilibrium. When an AI Overview synthesises information from multiple sources and presents a coherent answer directly on the search results page, the user's need to visit the underlying sites diminishes. Publishers lose traffic; Google retains the user and the associated ad inventory.
Mehta's opinion acknowledges this tension but declines to resolve it through antitrust enforcement. The absence of an explicit agreement means there is no breach of contract. The lack of coercion, in the court's view, means there is no antitrust violation. What remains is a business relationship that one party has unilaterally altered, leaving the other with little recourse under current law.
Policy Levers Google Controls
Google offers publishers a mechanism to opt out of AI Overviews while remaining in traditional search results. The tool relies on a specific value in the robots.txt file that webmasters place on their servers. This technical option exists, but its practical value is debatable. As AI Mode and AI Overviews occupy increasing portions of the search results page, particularly on mobile devices, the distinction between being excluded from AI features and being excluded from meaningful visibility begins to blur.
Other publishers beyond Penske and Chegg have raised similar concerns. News organisations, recipe sites, and reference platforms have all reported traffic declines coinciding with the expansion of AI-powered search features. Some have negotiated licensing agreements with Google and other AI developers; others have opted out entirely. The litigation dismissed by Mehta represented an attempt to use competition law to force a different outcome.
That strategy has now failed in court, but the underlying economic conflict persists. Google controls both the distribution channel (search) and increasingly the content presentation layer (AI summaries). Publishers depend on that distribution but have limited leverage to negotiate terms. The technical opt-out Google provides is a unilateral policy, not a negotiated settlement, and the company retains the ability to modify it.
What the Ruling Does Not Settle
Mehta's decision resolves the specific antitrust claims brought by Penske and Chegg. It does not address several adjacent questions that remain live in other jurisdictions and regulatory contexts. Copyright law, for instance, operates on different principles than antitrust law. Whether using published content to generate AI summaries constitutes fair use or infringement is a question multiple courts are currently considering in separate litigation involving OpenAI, Microsoft, and other developers.
Regulatory bodies outside the judiciary are also examining the intersection of AI and search. The European Commission has opened inquiries into whether Google's integration of generative features violates the Digital Markets Act, which imposes specific obligations on platforms designated as gatekeepers. Those proceedings operate under a different legal framework and may produce different outcomes.
The broader policy debate extends beyond litigation. If generative search features fundamentally alter the economics of online publishing, the question becomes whether existing legal structures are adequate to address the shift. Antitrust law, as Mehta's ruling demonstrates, requires proof of specific harms and anticompetitive conduct. It is not designed to preserve incumbent business models or enforce informal expectations.
The Compact Unravels
The dismissal of these lawsuits leaves publishers in a weaker position than they occupied before generative AI entered search. The implicit traffic compact that structured the web for two decades has been rewritten by one party, and the courts have declined to intervene. Publishers retain the option to exclude their content from AI features, but doing so carries its own costs in terms of visibility and relevance.
For Google, the ruling removes one legal obstacle but does not resolve the broader tension. The company faces ongoing scrutiny from regulators, potential copyright challenges, and a deteriorating relationship with the publishing industry that supplies much of the content its AI features rely upon. The technical and legal mechanisms Google has deployed may satisfy antitrust standards, but they have not produced a stable equilibrium.
The web is entering a phase in which the distribution layer and the content layer are no longer clearly separated. Search engines are becoming answer engines, and the traffic that once flowed to publishers is increasingly retained by the platforms that index them. Mehta's ruling confirms that this transformation is not, in itself, a violation of competition law. Whether it is sustainable remains an open question.
