Media companies have filed so many lawsuits against AI companies over the past two years that the act has become routine. When I report on these in The Media Copilot newsletter, they’re often digest items, adding to the pile of publishers who want fair compensation for the content AI labs have ingested to create large language models (LLMs). There are so many that elaborate infographics are required to keep track of them all.

Penske Media’s copyright lawsuit, however, is anything but typical, and that’s because of its choice of target. The Rolling Stone publisher is going after Google.

Google is in many ways the big fish in the AI world. It’s true that more people use OpenAI’s ChatGPT than they do Google Gemini, but Google has the distinction of being both a frontier AI lab and the current owner-operator of the primary way people get information on the internet. If you count AI Overviews in search, Google is arguably bringing AI to more people than anyone on the planet, period.

And because of that, the media is starting to feel serious pain. As more Google searches result in answers instead of 10 blue links, people aren’t clicking like they used to, and the referral traffic that publishers depend on is drying up. To add insult to injury, Google uses a single crawler to index websites for both regular search and AI, forcing publishers to allow it to harvest their content for Overviews even while search is sending them fewer and fewer referrals.

The dam is bursting

Still, nobody wants that referral number to go to zero, so virtually no one in the publishing world has dared poke the big G. When News Corp sued Perplexity last year, there weren’t that many AI search engines yet, but there was at least one more: Google. Everything News Corp cited in its complaint—that Perplexity ingested content, used it to build a competing product in its AI answers, and didn’t bother to pursue content licensing—could equally apply to Google, yet no second lawsuit was filed.

Will that change now that Penske Media has made its move? And I don’t mean just for News Corp—if the publisher of Variety, The Hollywood Reporter, and Billboard has looked at the numbers and decided legal action is worth the cost and risk, how many others are coming to the same conclusion? On the other hand, this might play out more like The New York Times v. OpenAI, with other publishers mostly watching from the sidelines, reasoning they could play it safe and still benefit if there’s a decision favorable to the media.

We’ll know the answer to that soon enough, and yes, technically Penske Media isn’t the first content provider to sue Google over AI Overviews; Chegg, an online educator, filed suit against the $3 trillion tech giant back in February. But Penske’s lawsuit does something else: It exposes Google’s claim that AI Overviews send “higher quality” traffic to publishers as a meaningless consolation prize.

Last fall, when the effects of Overviews were just beginning, Google said that people who click on links in the summaries were more intentional users, and thus more likely to engage and even transact with the sites they click on. And that may be true, but it’s telling that Google still hasn’t provided concrete data to back up that claim—specifically how much Overviews reduce click-through rates overall.

And that’s OK, because now we have third-party data on the click-through rates of AI answer engines, and they’re dismal. Pew Research, Similarweb, and TollBit have compared the crawl-to-referral ratios for AI engines to search engines, looking at how many users they actually send versus how often they scrape content (which usually is indicative of a search), and they don’t look good. Click-through from AI search is 90% lower than regular search, per TollBit.

In other words, the drop-off in referrals from AI summaries is so massive that, in almost all cases, the “higher-quality” visitors couldn’t possibly make up for the loss in business from the lost search referral traffic. And remember: Google doesn’t separate its search and AI bots—you either let both crawl your site, or you get nothing.

Google Nero?

Google is rumored to have begun talks with publishers to license content for its AI, which would be a massive shift, and possibly quell the impulse for other publishers to join Penske in the legal arena. However, that would turn Google AI search results into another OpenAI: summaries that included content from publishers big enough to have leverage, leaving smaller players with little recourse but to continue to provide their content to Google for free.

If the two Google crawlers were separate, things would be different. Publishers of any size could opt-out of AI Overviews while maintaining their position in regular search. The fact that Google is considering paying publishers for AI is evidence that the two technologies are governed by different business models. Forcing a site to give up content for both is like a gas station that forces you to get a car wash before it’ll fill up your tank.

Over the past decade, the publishing world has woken up to the predations of Big Tech. The search/social era didn’t turn out well for the industry, and now it wants to choose a different path with AI. Except now tech companies like Google have so much leverage over distribution, the choice has been taken away. Can a lawsuit get it back? Maybe, maybe not. But at least it shows the media is no longer afraid to try.