Skip to content
World clockEU--:--UK--:--USA--:--CN--:--PLDEFRIT中文EN

portal about AI and technologyevents · analysis · interviews · technical background

Search
LIVE
›

Google's AI Overviews survive antitrust suits as publisher traffic keeps falling

A US judge has dismissed antitrust lawsuits from Chegg and Penske Media over Google's AI Overviews. The judge ruled the feature did not drain the audience and revenue the publishers claimed. New data shows search referrals to North American publishers falling to 12%.

Media & internetExplainerRachel NwosuPublished: 3 October 20264 min readSources 9
Google's AI Overviews survive antitrust suits as publisher traffic keeps falling

The ruling, reported by Quartz and Technology Org on 2 October, throws out claims that Google's AI-generated summaries at the top of search results illegally diverted readers and ad money. The publishers said those summaries were built on their work. The judge had previously found Google to be a search monopolist in a separate case. That detail makes the outcome sting for publishers.

For anyone who runs a site that depends on Google referrals, the legal route to protecting that traffic is now much narrower. The court accepted Google's position that AI Overviews did not cause a measurable loss of audience or revenue. Google has made that same claim publicly for two years.

Independent measurement tells a different story. Chartbeat data reported on 20 October put search share at 12% for North American publishers. That figure describes how much of their audience arrives via search engines rather than direct or social channels. That is the number to watch, not the courtroom verdict. It also sits awkwardly beside the finding the judge accepted. A 12% search share measures dependence, not harm, and the two numbers answer different questions.

The timing is awkward for the complainants. Chegg and Penske Media argued in part that AI summaries substitute for the click. The summaries answer a query before a reader ever reaches the source. The judge found no audience and revenue drain, according to Technology Org's account of the dismissal. Whether that finding survives appeal is unknown. Neither company has said publicly whether it will refile or challenge the decision.

What the traffic data shows

Reports collected over the past several months describe a steady erosion rather than a cliff. One analysis from 1 November found that AI search is cutting publisher traffic as search engines lose ground overall. The loss comes from two directions at once. Fewer people start on a search engine, and fewer of those who do click through. A separate report from 25 October said Google's AI search changes had upended online publishing. It put no figure on the disruption.

An earlier report, dated 8 April, said AI bot traffic had surged 300%, hitting publishers hardest. That number measures crawler activity, not human readers. The two should not be conflated. More bots reading a page does not mean more people see it, and it does not mean more ad impressions are sold. Publishers have spent the year trying to model what a future with significantly less Google search traffic looks like, according to a 1 June analysis. The modelling work itself shows how seriously the industry now takes the decline.

Google, for its part, has been paying some publishers for AI content. Buttondown flagged that development on 2 October. The terms of those arrangements have not been made public. It is not clear how many publishers are covered or how the payments are calculated. Neither Google nor the publishers involved have disclosed figures.

"Are AI Overviews killing websites? A court just sided with Google," Android Authority wrote on 2 October, framing the ruling as a legal win that leaves the underlying question open.

That framing is fair. The court decided a specific antitrust claim on a specific evidentiary record. It did not decide whether AI summaries are good or bad for the open web. It did not set a rule that other publishers can rely on in future cases, at least not on the reporting available.

Where the pressure moves next

With the courtroom route narrowed, attention shifts to the two levers publishers still hold: licensing deals and product changes. The licensing route is opaque. The product route is slow. Meanwhile the measurement firms keep publishing the same direction of travel.

Search Engine Land reported on 2 October that Google had described how its search quality raters evaluate main content. The ranking system still rewards original work, even as the results page increasingly answers questions without sending anyone to it. Google Ads also loosened business name rules for some advertisers the same day. That change has no direct bearing on publishers. It is part of the same week of adjustments to how Google's surfaces present information.

The most recent signal is the weakest for the industry. The 1 November analysis of AI search cutting publisher traffic arrived after the ruling and after the Chartbeat figure. It points the same way: search is a smaller door than it used to be. Publishers who built their businesses on that door are now modelling, licensing and litigating at the same time. Only one of those three just got harder.

Comments 0

Sources

9
  1. 01Palantir's fondness for French food cooked up tech's latest fad – forward-deployed engineersEN
  2. 02Buying used CPUs can expose users to existing bans from anti-cheat enginesEN
  3. 03There are only 5,000 elite software engineers in the world (2025)EN
  4. 04Former SR-71 Engineer Talks NASA's Blackbird Revival ProgramEN
  5. 05Capcom Plans to Transform the RE Engine into an AI-Generation Game EngineEN
  6. 06The Rise of Overfit Inference EnginesEN
  7. 07The search for Amelia Earhart's plane beginsEN
  8. 08AI Engineering from ScratchEN
  9. 09Taming Coding Agents Using Chemical Engineering PrinciplesEN

All figures and quotations in this text come from the sources listed below.

Content prepared by the editorial team with AI assistance.

Rachel Nwosu

Rachel Nwosu

AI, models and technology

Rachel Nwosu covers AI, models and technology for FLASH24, working from public model documentation, benchmark releases and repository histories rather than press summaries, and she skips announcements that arrive without reproducible numbers. She checks training-data claims against dataset cards and reruns reported metrics where code is available. She spends much of her week interviewing researchers and engineers, tracking model launch calendars, and comparing vendor benchmarks with independent evaluations. Outside the desk she runs 3D printers, restores old computers, and tests how models learn from internet junk. She does not publish benchmark figures she cannot trace to a source.

Newsroom →

Comments

0
  1. No comments yet — be the first.

Write a comment

Comments are public. We do not publish abuse, spam or advertising.