Google Search is no longer just a list of links. With AI Overviews now appearing at the top of many results pages, users are increasingly seeing a generated answer before they decide whether to visit another website.
A Pew Research Center analysis, based on data from 900 users of the Ipsos KnowledgePanel collected in March 2025, suggests that this change is already affecting search behavior. Searches that include AI Overviews were linked to a lower rate of clicks to other sites, even as Google argues that its AI features create new opportunities for people to connect with websites.
What Pew found in Google search behavior
The central finding is straightforward: when Google showed an AI Overview, users clicked through to other websites less often.
According to Pew, searches without an AI answer produced a click rate of 15 percent. When the results page included an AI Overview, clicks to other sites fell to 8 percent. That is a major difference for publishers, businesses, and any site that depends on search traffic.
The study also examined clicks on the sources shown inside the AI Overviews themselves. Google has said that people click on links cited in AI Overviews, but Pew found that just 1 percent of AI Overviews led to a click on a source.
The most common sources in these AI answers were Wikipedia, YouTube, and Reddit. Together, they accounted for 15 percent of all AI sources.
Why AI Overviews matter for publishers
For web publishers, the issue is not only whether Google still sends traffic overall. The more immediate concern is that a search results page can now answer a user before that user reaches the reporting, analysis, or reference material behind the answer.
That changes the basic exchange that has long shaped search. Websites publish information, Google organizes and displays it, and users choose which result to open. AI Overviews place a generated response above that decision point.
If fewer users click through, publishers may lose visits even when their work helps inform the search result. This is especially important for sites that rely on search traffic to support editorial work, product information, tutorials, reviews, or other web content.
Pew’s numbers also raise a separate visibility problem. Even when an AI Overview cites sources, the source links may not receive much user attention. A citation inside an AI answer is not the same as a visit to the original page.
More searches are getting AI answers
The effect matters more because AI Overviews are not a rare feature. Google began testing them as the “search generative experience” in May 2023. A year later, they became an official part of the search engine results page.
Pew says about 1 in 5 searches now have AI Overviews. The analysis also found that longer searches are more likely to trigger an AI answer.
Question-style searches appear especially likely to produce an AI Overview. Pew reported that 60 percent of questions and 36 percent of full-sentence searches are answered by the AI.
That pattern is important because users often turn to search engines when they need explanations, definitions, comparisons, or practical guidance. Those are exactly the types of queries where a generated answer may reduce the need, or the perceived need, to open another result.
The information risk behind fewer clicks
Lower click-through rates are not only a business issue for publishers. They also affect how people check information.
The source article notes that Google users are more likely to end their browsing session after seeing an AI Overview. If the generated answer satisfies the user, the search may stop there.
That behavior can be convenient, but it can also narrow the user’s view. Generative AI systems can produce incorrect information through “hallucinations.” If users do not open sources, compare results, or continue investigating, they may leave a search with an answer that appears complete but is wrong.
This is a different kind of search experience. Instead of scanning multiple results and judging the source, users may rely on a single summarized response. The tradeoff is speed versus verification.
Google rejects Pew’s conclusion
Google disagrees with the findings and the way the research was conducted. The company’s statement said: “People are gravitating to AI-powered experiences, and AI features in Search enable people to ask even more questions, creating new opportunities for people to connect with websites. This study uses a flawed methodology and skewed queryset that is not representative of Search traffic. We consistently direct billions of clicks to websites daily and have not observed significant drops in aggregate web traffic as is being suggested.”
The disagreement reflects a larger conflict over the future of search. Google frames AI-powered search as a way to help users ask more questions and reach websites in new ways. Pew’s analysis points to fewer clicks when AI answers appear.
Both claims focus on the same shift: AI is changing the path between a search query and the wider web. For readers, that means the first answer may arrive faster. For publishers, it may mean fewer opportunities to receive a visit. For the search ecosystem, it raises a direct question about how much of the web users will still see once the answer is summarized at the top of the page.