Google's AI Overviews are decimating referral traffic to publisher websites. Pew Research tracking 900 U.S. adults found that when Google displays an AI summary in search results, users click traditional links just 8% of the time, down from 15% when no summary appears. Links cited within the AI answers themselves get clicked roughly 1% of the time.
The data is grim for content creators. Chartbeat reports that page views from Google Search dropped 34% across its publisher network between December 2024 and December 2025. Sites that depend on search traffic discovery face a structural problem: Google now answers questions directly, extracting information from multiple sources and presenting it without requiring users to visit the originating pages.
The core issue is architectural. Most websites are optimized for the old web, where search traffic funneled readers through the homepage and into deeper content. That funnel is collapsing. AI systems cite sources buried in search results but present summaries prominently, incentivizing users to stay in Google rather than click through.
Publishers face a compounding problem. They still need to rank well to be cited by AI systems. But being cited no longer drives meaningful traffic. Google's system extracts value from their reporting, surfaces it in the search interface, and keeps users on Google's platform. The company collects advertising revenue while publishers lose clicks.
Some publishers are experimenting with paywalls and direct traffic channels to reduce search dependency. Others are pushing back against AI crawling. But Google's dominance in search and its control over AI Overviews create an asymmetrical power dynamic. Publishers need Google visibility to survive, but that visibility increasingly means serving Google's users rather than attracting them to their own sites.
This represents a fundamental shift in how information flows online. The web was built on clickthrough economics. AI search threatens that model without offering publishers an