AI Overviews: Why Google Broke the 10 Blue Links
For twenty-five years, the deal was simple. You typed a question into Google, you got ten blue links, you clicked one, and somewhere on the other end a publisher earned a sliver of ad revenue for having answered you. That deal is the reason the modern web i.e. blogs, news sites, recipe pages, SEO agencies, my own inbox exists in the shape it does. As of 2026, Google has quietly torn it up, and I don’t think it had much of a choice.
The number that explains everything
Start with this: AI-driven search interactions grew from under 10% of all search activity in 2023 to roughly 30% by 2026. ChatGPT alone now accounts for about 68% of AI chatbot traffic, and 42% of people say they’d rather ask an AI chatbot than run a traditional search when they’re doing multi-step research. Google still owns nearly 90% of the traditional search market, but “traditional search” is shrinking as a share of how people actually look things up.
That’s the context that gets left out of most of the outrage about AI Overviews. This wasn’t Google waking up one day and deciding to be generous with an AI summary out of the goodness of its heart. It was the company that invented the “10 blue links” business model watching a chunk of its users quietly migrate to interfaces that skip the links entirely. AI Overviews is a defensive product wearing an innovation costume.
What it actually did to the web
Whatever the motivation, the effect on the open web has been brutal, and the data backs that up more clearly every month.
Ahrefs’ research puts it starkly: pages that used to rank #1 and enjoy a 7.3% click-through rate are now seeing 1.6% when an AI Overview appears above them — a 58% collapse. Pew Research found that only 8% of users click through to a regular result when an AI Overview is present, versus 15% when it isn’t, and that a full 26% of people now end their search session entirely after reading the summary, never clicking anything at all. Digital Content Next reported member publishers losing a median of 10% of referral traffic in eight weeks, with some verticals down more than 75%.
Publishers aren’t taking it quietly. Penske Media (which owns Rolling Stone, Variety, and Billboard, among others) sued Google in February 2026, arguing the company “shattered the longstanding bargain” that built the open internet — take our content, index it, send us traffic in return. Their complaint frames what publishers now face as a forced choice: keep letting Google train and summarize your work for free, or get delisted and disappear from search entirely. Chegg has filed similar claims. It’s hard to read those and not sympathize — an entire industry built its business model around a handshake agreement that one side has now unilaterally rewritten.
The case Google would make — and it’s not nothing
Here’s where I’ll push back on my own instinct to just be angry about this, because the case for AI Overviews isn’t purely cynical.
A meaningful share of searches genuinely don’t need a website visit. “What’s the boiling point of water at altitude,” “how do I convert 200°C to Fahrenheit,” “who directed this movie” — these are the searches that built an entire cottage industry of content farms optimizing thin pages purely to sit above a fact nobody wanted to read 800 words to find. If AI Overviews kill that specific genre of SEO spam, I’m not going to mourn it. Pew and other researchers have also found that when people do click through after reading an AI Overview, they arrive having spent more time evaluating options first — session data from 846,000 searches shows users lingering longer, scrolling back to compare, engaging more deliberately rather than reflexively clicking the top result. That’s arguably a better-informed click, even if there are fewer of them.
And it’s not uniformly bad news for publishers, either — sources actually cited inside an AI Overview see about 35% more clicks than they otherwise would have gotten. The traffic isn’t just vanishing; it’s being redistributed toward whoever Google’s model decides is the authoritative source on a given topic, which is either an opportunity or a new, more opaque gatekeeper depending on how charitable you’re feeling.
Where I land
I don’t think “AI Overviews are good” or “AI Overviews are bad” is the right frame. What’s actually happening is a wealth transfer — attention and ad revenue moving from the open web toward whichever company controls the summarization layer — dressed up as a UX improvement. Some of the UX improvement is real. Some of the wealth transfer is genuinely destructive to an ecosystem of independent publishers who had no seat at the table when the rules changed.
What bothers me most isn’t the shift itself — search was always going to get more conversational, that was obvious years before AI Overviews shipped. It’s that Google gets to be judge, defendant, and referee all at once: it decides what counts as a citation-worthy source, it profits whether or not that source gets clicked, and publishers have no real alternative distribution channel that reaches anywhere near the same audience. That asymmetry, more than the AI itself, is the part actually worth being upset about. If there’s a regulatory fight worth watching in the next year, it’s not about whether AI Overviews should exist — it’s about whether the company running search should also get to unilaterally decide how the value it extracts gets shared with the web it was built on.
For those of us who write for a living, the practical takeaway is less philosophical: optimizing for “rank #1” is no longer the whole game. Optimizing to be the source Google’s model trusts enough to cite is the new one — and right now, almost nobody has that fully figured out.
Sources:
- Google’s AI Overviews and Publisher Traffic: 58% Click Decline and Antitrust Filing – ALM Corp
- 846,000 Google Searches Reveal How AI Overviews Are Changing User Behavior – Search Engine Journal
- Google AI Overviews Statistics 2026: 60+ Data Points – QuickSEO
- AI Search Market Share 2026 – Sedestral
- The Rise of AI Overviews: Opportunities and Challenges for Marketers – Envisionit
