AI web pages are multiplying, but is that why fewer people click through from search results?

This explores whether AI-written pages are crowding out human ones in search results while click-through traffic falls. The corpus suggests the two trends are real but don't fit together the way the question assumes. The bigger change isn't AI articles winning the rankings. It's AI answers sitting above the rankings, so fewer people scroll down to any article, human-written or not.

Start with supply. About a third of newly published websites were AI-generated or AI-assisted by mid-2025 How much of the internet is AI-generated now?. Those pages weren't less accurate, but they were more alike and more upbeat in tone. One estimate puts AI's share of new articles at about 50%, where it then stopped growing. A possible reason is that AI articles do badly in search, so producing more of them stopped paying off. Another is that newer models simply slip past the detectors. The source behind that 50% figure tests neither explanation and reports no traffic data Why did AI article share stop growing after 2025?. So the corpus can't tell you whether AI articles now dominate search results. Nobody here has measured that directly. It's also hard to measure at all, because people identify AI content at roughly chance level Can people reliably spot content made by AI?.

The traffic decline is much better documented, and the cause is the answer box, not the AI article. When Google showed an AI summary, users clicked a result 8% of the time instead of 15%, and more search sessions ended with no click at all Do AI summaries on Google reduce clicks to actual websites?. A natural experiment comparing English Wikipedia with its German and French editions put the loss from AI Overviews at about 5% of search referrals Does AI search summaries divert traffic away from Wikipedia?. News executives expect search referrals to fall 40–43% over three years Will AI Overviews reduce search referral traffic to publishers?. People haven't abandoned search, though. They mostly run AI chat alongside traditional search rather than instead of it Does generative AI chat actually replace traditional search?.

The less obvious point is where the two trends meet. A simulation found that when about two-thirds of a corpus is synthetic, over 80% of retrieved results come from synthetic sources. Answer accuracy stays high throughout, so nothing looks wrong. Retrieval tends to amplify synthetic content beyond its actual share, and the usual quality checks don't catch the loss of variety in sources Does synthetic content in search results hide ecosystem decay?. That's the risk to watch: AI summaries may increasingly be built from AI-written pages, and the reader sees neither. One argument holds that better search and curation can't fix this. The internet made existing knowledge easier to reach, but AI keeps generating new text without limit. Fixes would have to come at the production end, such as marking where content came from Why do search tools fail against AI generated content?.

Social media shows a similar pattern. There, AI posts collect likes and visibility without sparking conversation, so they push out human creators while hollowing out what made the platform worth using Does AI content displace human influencers on social media? Why do AI posts get likes without inviting conversation?. Search may be heading the same way. AI content may not need to top the rankings to reshape what people learn. It only needs to be what the answer box quietly draws on.