AI Search Is Eating the Open Web and Nobody Is Paying Attention

Here is a number that should terrify anyone who publishes content on the web: zero-click searches now account for nearly 65 percent of all Google searches. When you add AI Overviews, ChatGPT search, and Perplexity to the mix, the percentage of searches that result in a click to an external website is approaching an all-time low.

The web is being eaten from the inside. And the companies doing the eating are the same ones that built their empires on top of it.

This is not a prediction about the future. This is happening right now, in real time, and almost nobody outside of SEO circles is talking about it.

The Zero-Click Trap

For years, Google has been reducing the need to click through to websites. Featured snippets, knowledge panels, local packs, and now AI Overviews all serve the same purpose: keep the user on Google's property.

The logic is simple. More time on Google means more ad impressions. More ad impressions means more revenue. The user gets their answer faster, Google gets paid, and everyone is happy.

Except the websites that actually created the information being summarized.

When Google scrapes a recipe site to display a full recipe in a featured snippet, the recipe site loses a visitor. When an AI Overview summarizes an in-depth article into three paragraphs, the publisher loses a reader. When Perplexity generates a comprehensive answer citing twelve sources, none of those sources get a click.

The math is brutal. A study by SimilarWeb found that AI Overviews reduce clicks to top-ranking pages by 30 to 60 percent. For queries where AI Overviews appear, the click-through rate on organic results has collapsed.

This is not a temporary dip. It is a structural shift.

Why AI Search Makes It Worse

Traditional search at least pretended to send traffic to websites. The search results page was a directory. You picked a link, clicked, and visited a site.

AI search does not even pretend.

When you ask ChatGPT a question, it generates an answer. It might cite a source, but the citation is a footnote, not a destination. When Perplexity responds, it synthesizes multiple sources into a single coherent answer. The user has no reason to visit any of the original sources because the answer is already complete.

Google's AI Overviews are the most insidious because they combine the zero-click behavior of AI with Google's dominant market position. When Google decides to show an AI Overview, it pushes organic results below the fold. Users get their answer from the AI summary and leave.

The publishers who spent years building SEO equity, creating content, and earning rankings are watching their traffic vanish overnight.

The Content Creation Paradox

Here is the paradox at the heart of AI search: the AI needs content to learn from, but the AI's answers eliminate the incentive to create that content.

Why would a publisher spend hours researching and writing an in-depth guide if an AI is going to summarize it for free? Why invest in original reporting if an AI search engine will scrape the key facts and present them without sending a single visitor?

This is the tragedy of the commons, applied to the web. Every AI search engine benefits from the content created by publishers. But each individual AI search engine has no incentive to send traffic back. And when all of them behave this way, the content ecosystem collapses.

We are not there yet. Publishers are still creating content. But the trend lines are unmistakable. Traffic is declining. Revenue is shrinking. And the economic model that sustained the open web for twenty years is breaking down.

The GEO Gold Rush and Its Discontents

The response from the SEO industry to AI search has been to create a new discipline: Generative Engine Optimization, or GEO. The idea is to optimize content so that AI search engines cite your brand in their answers.

GEO is not wrong. If AI search is the new distribution channel, then brands need to figure out how to appear in it. But GEO is a band-aid on a deeper wound.

Even if your brand gets cited in an AI answer, you still do not get a click. You get brand visibility, which is valuable, but it does not pay the bills the way website traffic did. A mention in a Perplexity answer is worth a fraction of what a top-three Google ranking was worth two years ago.

The SEO industry is pivoting to GEO because it has to. But the pivot masks a fundamental problem: the economics of content creation on the web are deteriorating.

What Happens Next

Three scenarios are possible.

Scenario one: the walled garden accelerates. Publishers move behind paywalls and login screens to prevent AI scraping. The open web shrinks. AI search engines have less content to train on. Quality degrades. Everyone loses.

Scenario two: regulation forces compensation. Governments require AI search engines to compensate publishers for using their content, similar to the news media bargaining codes in Australia and Canada. This is possible but slow. By the time regulation catches up, many publishers will be gone.

Scenario three: a new economic model emerges. Micropayments, content licensing, or some other mechanism creates a sustainable way for publishers to earn money from AI-surfaced content. Several startups are working on this. None have cracked it yet.

The most likely outcome is some combination of all three. Some publishers will build walled gardens. Some will get regulatory protection. Some will find new models. And many will simply disappear.

The Real Cost of Free Answers

We have grown accustomed to free, instant answers. We ask a question, an AI responds, and we move on. It feels like magic.

But the answers are not free. They are paid for by the content creators whose work is being summarized, scraped, and synthesized without compensation. Every AI answer is a small extraction from the open web.

The cumulative effect of millions of extractions per second is a web that is slowly being hollowed out. The content is still there, but the traffic, the revenue, and the incentive to create new content are disappearing.

This is the real crisis of AI search. Not hallucinations. Not bias. Not even privacy. The crisis is that AI search is killing the economic engine that created the content it depends on.

And almost nobody is paying attention.

What Brands Should Do Now

If you are a brand or publisher, you need to adapt to this new reality. Here is what that means in practice.

First, diversify your traffic sources. Do not depend on Google or any single AI search engine for your audience. Build email lists, communities, and direct relationships with your users. Own your distribution.

Second, invest in GEO. Even though citations do not generate clicks, they generate brand awareness. Being cited by AI search engines builds trust and authority. It also increases the likelihood that users will seek you out directly.

Third, create content that AI cannot easily summarize. Original research, proprietary data, strong opinions, and unique perspectives are harder for AI to compress into a generic answer. This kind of content is also more valuable to your audience.

Fourth, monitor your AI visibility. Use tools that track how often your brand appears in AI-generated answers. What gets measured gets managed.

Fifth, consider licensing your content. Several AI companies are willing to pay for high-quality content. This is not a replacement for organic traffic, but it is a revenue stream that did not exist before.

The open web is not dead yet. But it is wounded. And the companies that wounded it are the ones we rely on to find things on it. The irony is thick enough to cut with a knife.

The question is whether we will notice before it is too late.