Observing large media corporations file federal lawsuits over issues they are unable to resolve in court has an almost poignant quality. Google is being sued by Penske Media, which owns Variety, Billboard, and Rolling Stone, over AI Overviews; similar claims have been made by Chegg, an educational content company. Perplexity has been targeted by Dow Jones and the New York Post. The list now includes Merriam-Webster and Encyclopedia Britannica. The complaint that permeates all of these cases is basically the same: AI search summaries are consuming our traffic, and someone has to foot the bill. It’s a serious grievance. Simply put, it’s not legal.
The main contention is that Google’s AI Overviews take content from publishers without paying them enough, present users with a condensed response at the top of the page, and do nothing. Google is not pleased with the data. Users clicked on a traditional result in just 8% of sessions when an AI summary appeared, compared to 15% when no summary was displayed, according to a Pew Research Center study that tracked over 68,000 Google searches. After AI Overviews were implemented, top-ranking page traffic decreased by about 34.5%, according to Ahrefs, an SEO analysis company. These figures show actual revenue losses. However, they do not constitute a cause of action under the law.
The most obvious aspect of the case is probably the copyright claims, which are also the most likely to be unsuccessful. A summary must typically replicate the original work in significant rather than transformative ways in order to violate copyright. Usually, AI Overviews doesn’t republish a single article in bulk. They combine data from several sources to create a fresh, useful solution. Under the fair use doctrine, courts have consistently viewed this type of synthesis as transformative, and transformative use is a potent defense. Although the claim that Perplexity “free rides” on the content of Merriam-Webster and Britannica may seem emotionally appealing, decades of case law indicate that creating something that is functionally distinct from the source material is precisely what fair use is intended to protect.
Key Information: AI Search Summary Lawsuits
| Field | Details |
|---|---|
| Core Legal Issue | Whether AI-generated search summaries constitute copyright infringement and antitrust abuse |
| Primary Defendant | Google (AI Overviews feature) |
| Key Plaintiffs | Penske Media (Rolling Stone, Billboard, Variety), Chegg Inc., Dow Jones, New York Post, Encyclopedia Britannica, Merriam-Webster |
| Competing Defendant | Perplexity AI (separate but related lawsuits) |
| Main Legal Theories | Copyright infringement, antitrust abuse, unfair competition |
| Primary Defense | Fair use doctrine; no legal right to search traffic |
| Key Research Finding | AI Overviews correlate with a 34.5% drop in organic click-throughs (Ahrefs); clicks on traditional results fell from 15% to 8% when summaries appear (Pew Research) |
| Legal Expert View | Summaries constitute transformative fair use; traffic loss is not an actionable legal harm |
| European Contrast | EU Digital Markets Act taking more aggressive stance; complaints filed with European Commission and UK CMA |
| Relevant Precedent | NYT v. OpenAI; Ross Intelligence v. Thomson Reuters; Australia and Canada news bargaining code experiments |
The antitrust angle is worth considering because it is more intriguing, if not significantly more feasible. Chegg contends that Google exploits its approximately 90% market share in search to essentially compel publishers to enter into a coercive agreement: consent to the use of your content in AI Overviews or risk having it completely removed from search results. There is a certain logic to that framing. However, courts have traditionally viewed modifications to search algorithms—even significant ones that completely destroy particular publishers—as product enhancements rather than anticompetitive behavior. Google’s counterargument, which claims that publishers can opt out using technical tools like robots.txt, ignores the fact that, for the most part, opting out of Google entails opting out of the internet. However, demonstrating that Google’s actions constitute anticompetitive harm rather than just aggressive competition is a very high bar.
Beneath all of this is a deeper structural issue. The recognition of a right to traffic is what these lawsuits are really aiming for, albeit indirectly. a click-based property interest. Furthermore, American law just does not recognize that right. According to the courts, search rankings are a type of platform expression. A search engine is not required by law to direct users to any particular location. The platform has always had the final say over how to present information, what to summarize, and what to link to. Google’s referral traffic was never guaranteed to publishers. It was always able to be stopped by Google. The underlying legal reality remains unchanged even though AI Overviews make stopping that traffic seem more methodical and deliberate.
The historical parallels are instructive, and this is not new ground. Ten years ago, governments in Australia, Canada, and Europe attempted to use ancillary copyright laws and news bargaining codes to legislate their way to a solution. For publishers, the outcomes were generally dire. For a short while, it was successful when France fined Google to compel talks. Google News simply shut down completely in Spain in 2014 as a result of the link tax. Meta completely removed news from Facebook as a result of Canada’s Online News Act. These experiments repeatedly showed that publishers required the platforms far more than the platforms required them, and that legal pressure could lead to exits rather than checks.
Since at least the earlier disputes involved the actual reproduction and display of content, the current wave of litigation may be even more precarious than those legislative initiatives. Articles are not republished by AI Overviews. They are creating something fresh by synthesizing various sources and delivering it to a user who might not have clicked through to any specific publisher in the first place. The copyright claim is undermined if summaries are not copies. Furthermore, it becomes extremely difficult to maintain the entire structure built on top of that claim if they are not copies.
There’s a sense that the real struggle isn’t taking place in the Southern District of New York courts at all as you watch all of this. The question of how journalism, reference publishing, and instructional content are financed in a world where AI can approximate their output on demand is one area where this is occurring in the economics of content creation itself. That is a real and significant crisis. Simply put, federal copyright law was not intended to handle it. The lawsuits will proceed as planned. Some will withstand narrow-ground dismissal motions. Some might quietly settle in. However, Variety’s complaint won’t stop the AI-driven search trend.
When the outdated traffic-to-advertising model is no longer viable, the more candid discussion—the one that genuinely has a chance of yielding something beneficial—focuses on what a functioning information economy looks like. With the Digital Markets Act, Europe is at least making an effort to have that discussion, with regulators scrutinizing Google’s actions more closely than American courts are likely to. It’s genuinely unclear if that results in a feasible solution or just another iteration of the platform walking away.
As of right now, Google continues to display AI summaries at the top of the page, the publishers are in court, and the attorneys are billing. What is filed in federal court is unlikely to significantly alter that image.
