Is the Dead Internet Theory Coming True?

AI summaries are driving a fast shift toward “zero click” search—where users get answers without visiting the original websites—and that threatens the business model of search-dependent industries, especially news. Google and other search engines long determined which sites received traffic, and traffic translated into revenue. Now AI systems increasingly deliver the needed information directly in the results page, cutting off the clicks that publishers rely on. Multiple analyses of Google data suggest this is already happening at scale: a Similarweb-based study estimated that in 2024 roughly 60% of Google searches in the United States and the European Union were “zero click,” up from about 26% in 2022. Other reporting cited that around 80% of consumers rely on zero click results for at least 40% of their searches. The impact is uneven—topics with Google AI Overviews see particularly large drops, with one figure putting declines around 25%—and some media outlets report severe losses. WorthDotCom said it lost nearly one third of its traffic after Google’s AI summaries were introduced, and a Daily Mail executive claimed click-through rates can fall by 80–90% when an AI Overview appears above a story.

So far, overall search usage hasn’t collapsed: people still search. But the trajectory points toward a further change as AI agents become more capable and more widely available, increasingly handling searches on users’ behalf. For now, many AI searches still rely on traditional search engines because those services have already indexed much of the web. Yet AI companies are also expanding their own coverage through crawlers that continuously scan the internet and feed content into their systems. The more they index directly, the less they need to depend on existing search engines—an evolution that could shrink the conventional web’s role as an intermediary.

That shift is already reshaping how websites respond. Some site owners try to block AI bots, while others “cooperate” by presenting content in formats easier for automated extraction. The likely end state is a web optimized for machine readability rather than human presentation: simpler layouts, more extractable text and images, and less emphasis on visual design.

Researchers have even framed the emerging landscape as an “agent attention economy,” where advertising and ranking systems target not direct persuasion of users, but control over which options an AI agent even considers. Tim Berners-Lee, by contrast, has suggested an “intention economy,” where users state goals and agents execute them—though the practical outcome may still concentrate power in the intermediaries that decide what gets surfaced.

The fallout is spilling into courtrooms. Chegg sued Google, arguing AI Overviews use its content to answer questions without sending students to Chegg’s site, harming traffic and revenue. In Europe, publishers filed an antitrust complaint alleging Google’s AI summaries divert money while still being trained on their material.

For humans, the “dead internet theory” is often treated as a conspiracy claim that most online interactions are bot-driven. Yet the current direction—more outsourcing of tasks to AI agents that can interact with services and each other—makes that scenario feel less far-fetched. The core worry is not that people stop using the internet, but that the internet’s economic center of gravity moves away from publishers and toward AI intermediaries that answer, rank, and monetize without routing users to the sources.