By Stuart Kerr, Technology Correspondent, LiveAIWire
AI Mode UK launch data confirms what publishers had feared since Google rolled the feature out on 28 July 2025: fewer clicks are reaching outside websites. Google’s own product blog, written by Hema Budaraju, VP of Product Management for Search, framed the launch as giving UK users a more powerful way to ask complex, multi-part questions and dig deeper into topics through follow-up prompts. Independent research published the same week the UK rollout began tells a more complicated story about what that shift means for the sites AI Mode draws its answers from.
A Pew Research Center analysis of 900 U.S. adults’ browsing data, covering the same period Google’s AI features were expanding, found that when a Google search produced an AI summary, users clicked through to a traditional search result in just 8 percent of visits. Without an AI summary present, that click rate nearly doubled, to 15 percent. Users clicked on a link inside the AI summary itself in just 1 percent of visits. The data covers AI Overviews specifically rather than AI Mode, but the underlying mechanism, an AI-generated answer displacing the need to click through, is the same one now being extended more aggressively in AI Mode’s fuller search interface.
What Google Says AI Mode Actually Does
Google’s official framing rests on a specific technical claim. AI Mode uses what the company calls a query fan-out technique, breaking a complex question into multiple sub-topics and running several searches simultaneously, then surfacing a wider and more varied set of web sources than a single traditional search would. Budaraju’s post states that early users are asking questions two to three times longer than typical search queries, and that Google has observed AI Overviews sending users to a greater diversity of websites for complex questions, with those clicks tending to be higher quality, meaning users spend more time on the sites they do visit.
What This Means for Anyone Publishing Content Online
For content creators, educators, and small publishers, the practical calculus has shifted rather than disappeared. Pew’s data shows Wikipedia, YouTube, and Reddit remain the most commonly cited sources in both AI summaries and standard results, together accounting for roughly 15 to 17 percent of links either way, meaning large, well-established sources retain visibility even as overall click volume falls. Government and educational sites appear notably more often in AI summaries than in standard search results, a detail worth knowing if your content sits in that category. The practical response for smaller publishers is less about opposing AI Mode and more about ensuring content is structured clearly enough to be one of the sources an AI summary actually cites, since being cited without being clicked is still a form of visibility, even if it does not convert to direct traffic the way it once did.
The Regulatory Pressure Building Around This
The UK’s Competition and Markets Authority has been considering whether to designate Google a strategic market status provider under new digital markets legislation, a status that could bring specific transparency and fair dealing obligations around exactly this kind of AI summarisation. Separately, media organisations across Europe have flagged tension between AI summarisation and the EU’s Copyright Directive, even though the UK itself is no longer bound by EU copyright law post-Brexit. Google’s position throughout has been that its own data shows overall web traffic from search remaining broadly stable, a claim the company has repeated even as it has disputed the methodology behind Pew’s specific findings.
Where the Real Disagreement Sits
The genuine dispute is not really about whether click-through rates on individual search result pages have fallen when an AI summary appears, since Pew’s tracked browsing data on that specific point is hard to argue with. The disagreement is about what that means in aggregate. Google’s position is that AI features are creating entirely new categories of search behaviour, net new queries that would not have happened at all without a conversational AI interface, rather than simply cannibalising clicks that would otherwise have gone to publishers. Publishers’ position is that regardless of whether some searches are net new, the ones that used to reliably send a click to their site increasingly do not, and that distinction matters more to a newsroom or education platform’s revenue than the aggregate search volume statistic does.
As explored in our coverage of AI’s invisible infrastructure and the silent bias questions embedded in automated systems, tools that decide which sources to surface and which to leave uncited carry an editorial weight that is easy to overlook precisely because the tool feels like plain search rather than a curated product. AI Mode’s UK rollout does not resolve that tension. It intensifies it, extending a fuller, more capable version of the same summarisation behaviour Pew’s data has already shown reduces the traffic reaching the sites an AI answer draws from. Whether UK regulators treat that as a competition question, a copyright question, or simply an acceptable cost of a better search experience will shape how this plays out across the rest of Europe as AI Mode continues its rollout.
About the Author
Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, emerging technology, and their impact on business, society, and everyday life. LiveAIWire publishes original AI journalism every weekday at liveaiwire.com.