Google’s AI Search: Beyond GEO, Facing the Garbage SERP Problem
▼ Summary
– Google representatives have debunked the common SEO advice to artificially “chunk” content for AI search, stating their systems are designed to access content as humans do.
– The core change in SEO for AI search is the shift from optimizing for a single keyword to addressing “query fan-out,” where one user query generates multiple related questions in a long-form answer.
– The article criticizes Google’s AI search for frequently surfacing low-quality, non-expert sources while hiding reputable, expert publications under secondary tabs like “More > News.”
– A key concern raised is that AI search’s query fan-out drastically reduces referral traffic to websites by answering multiple questions within Google’s interface, ranking only a few pages for many queries.
– Despite the new AI search interface, Google advises publishers to continue focusing on creating foundational, high-quality content for humans, as optimizing for current AI systems may not be sustainable.
The landscape of search is undergoing a fundamental transformation, moving beyond simple keyword matching to a model where Google’s AI aims to understand and answer the broader intent behind a query. This shift from returning a single link to generating comprehensive, long-form answers changes how content is discovered and valued. While the core infrastructure remains Google Search, the strategy for reaching audiences must evolve, as the old paradigm of optimizing for one keyword per page is effectively splintered by what’s known as query fan-out.
In a recent podcast, Google’s Danny Sullivan and John Mueller addressed how publishers should adapt. A common piece of advice circulating in SEO circles has been to restructure content into distinct “chunks,” based on the assumption that AI systems process information this way. Sullivan explicitly warned against this tactic. He emphasized that Google’s systems are designed to access and understand content as a human reader would, through proper HTML structure, headings, and logical organization. Artificially breaking content into bite-sized pieces specifically for AI is not only unnecessary but counter to Google’s longstanding guidance to create for people first.
Sullivan further cautioned that over-optimizing for current machine behaviors is a short-sighted strategy. As AI systems improve, they increasingly reward genuinely helpful, human-centric content. Those who focus on foundational quality and expertise are more likely to see sustained success than those chasing technical tricks meant to please an algorithm’s perceived preferences.
However, this guidance sidesteps a more pressing issue for publishers: the tangible impact on website traffic. The query fan-out phenomenon means Google’s AI answers now pull from multiple sources to address several related questions at once, drastically altering traditional referral patterns. This compounds a growing problem observed by many in the industry, the perceived decline in the quality of search results themselves.
There is a concerning trend where expert, authoritative content appears to be buried, often hidden behind “More” or “News” tabs, while the primary AI-generated answers cite sources of questionable expertise. For instance, a search for styling advice might reference an abandoned blog from 2018, a post on a business social network, or an article on a sneaker retailer’s site, rather than established fashion publications. This isn’t about cherry-picking bad examples; it’s a frequent experience that raises questions about how Google’s AI evaluates authority and trustworthiness.
The core issue transcends debates over whether to call it GEO, AEO, or traditional SEO. The practical result is that many high-quality sites are experiencing a significant drop in traffic, not due to a lack of expertise, but because of how the new search paradigm surfaces information. The joy of discovering a new, credible website through search seems diminished, replaced by AI summaries that often draw from marginal sources.
Ultimately, while Google advises creating excellent content for humans, which remains sound advice, publishers are left grappling with a system that doesn’t always seem to reward that excellence in its most visible outputs. The call from many in the community is for a renewed focus on surfacing truly authoritative sources, ensuring that the shift to AI-powered answers enhances, rather than undermines, the quality and reliability of search.
(Source: Search Engine Journal)
