Pew data shows AI summaries are now mainstream in search

Agencies finally have a clean number to put in front of skeptical clients: 60% of U.S. adults say they have ever read AI summaries at the top of search results. Pew Research Center’s June 2026 survey turns AI overviews from a fringe curiosity into everyday search behavior, and that makes the case for changing how search teams forecast clicks, measure visibility and explain traffic loss.

The survey covered 5,119 U.S. adults in the American Trends Panel and was fielded Feb. 17-23, 2026. Alongside the 60% who said they had read AI summaries, 30% said they had not and 10% were not sure, a split that matters because it suggests some people are encountering AI-generated answers without clearly recognizing them. Pew’s earlier October 2025 data showed the same pattern building fast, with 65% saying they at least sometimes came across AI summaries in search results and 45% saying they saw them extremely often or often.

AI chatbots are no longer niche either. Pew said about half of U.S. adults now use them, up from roughly one-third in 2024, and about one in four uses them daily. Searching for information was the most common use case, ahead of entertainment, image generation, medical advice and emotional support. ChatGPT remained the dominant product by a wide margin, with 44% of adults reporting use, followed by Gemini at 24%, Microsoft Copilot at 17%, Meta AI at 14%, Grok at 8%, Claude at 6% and Character.ai at 3%.

For agencies, the bigger warning is not just adoption, but click behavior. Pew reported in July 2025 that Google users were less likely to click result links when an AI summary appeared, and that they very rarely clicked the cited sources. That is the kind of datapoint that should change the conversation around CTR forecasting, because a classic blue-link ranking no longer captures all the places a brand can show up, or all the places a user can get an answer.

Pew’s usefulness data adds another layer. In the October 2025 survey, only 20% said AI summaries were extremely or very useful, while 52% called them somewhat useful and 28% said they were not too or not at all useful. Americans’ views on AI’s impact still tilt negative even as usage rises, which is exactly why agencies need to widen search reporting beyond rankings and traditional organic clicks. AI summaries, chatbot answers and standard search results are now competing and overlapping surfaces, and the businesses that measure only one of them will keep understating both exposure and risk.