Google has recently attempted to lower the temperature around GEO, AEO, and “AI SEO,” advising publishers not to overthink how they structure content for large language models. But in focusing on tactical advice like avoiding content “chunking,” Google may be missing a more urgent conversation: the declining quality and credibility of AI-powered search results.
SEO Has Changed—Even If the Foundations Haven’t
Google is correct when it says that the underlying infrastructure of Search has not been replaced. SEO fundamentals still matter. However, the way answers are presented has shifted dramatically.
AI-driven results now generate long-form responses that address multiple related questions at once. This query fan-out breaks the long-standing model of optimizing one page for one query. Visibility is no longer about ranking—it’s about being selected as a source within an aggregated answer.
That shift changes referral dynamics, traffic distribution, and which sites benefit from search exposure.
The “Chunking” Debate Misses the Point
In Google’s Search Off the Record podcast, Danny Sullivan discouraged publishers from restructuring content into artificial “chunks” designed for LLM consumption. That guidance is reasonable.
Well-structured web content already uses headings, lists, and semantic markup that both humans and machines can understand. Creating fragmented versions of content purely for AI systems risks harming readability and long-term value.
But debating chunking versus not chunking sidesteps the bigger issue.
AI Search Is Struggling With Source Quality
The more pressing concern for publishers and marketers is not how to format content—but which sources Google’s AI systems are choosing to elevate.
Across many queries, AI Mode surfaces content from:
-
Outdated or abandoned blogs
-
User-generated platforms with limited editorial oversight
-
Commercial sites lacking subject-matter authority
Meanwhile, established expert publications are often buried behind secondary navigation layers, such as the “More” or “News” tabs. This creates a paradox where authoritative content exists—but is no longer the default.
For users, this reduces trust. For publishers, it erodes the incentive to invest in high-quality expertise.
The Impact on Expertise and Discovery
Historically, Google rewarded expertise by sending traffic to sites that demonstrated authority, depth, and editorial standards. AI-driven SERPs increasingly blur those signals.
When AI systems prioritize convenience over credibility, the result is a flattened information landscape. Discovery suffers. So does differentiation.
This is not just an SEO problem—it’s a quality problem. If expert voices are consistently sidelined, the ecosystem that produces reliable content weakens.
GEO vs. SEO Is the Wrong Debate
Whether the industry calls this GEO, AEO, or simply SEO is largely irrelevant. Labels do not change outcomes.
What matters is that AI-driven search currently struggles to consistently surface the best sources. Until that is addressed, tactical guidance about optimization will feel disconnected from publisher reality.
What Google Needs to Address
If Google wants AI search to succeed long-term, it must confront several issues head-on:
-
How expert content is identified and prioritized
-
How AI systems balance ease of synthesis with source credibility
-
How referral traffic supports the sustainability of high-quality publishing
Without meaningful improvement, AI search risks becoming efficient—but unreliable.
The Bigger Picture
AI-powered answers are not inherently the problem. The problem is when convenience replaces judgment.
Search has always been about trust. If AI SERPs consistently elevate low-quality sources while hiding authoritative ones, users and publishers alike will lose confidence.
The future of search depends less on how content is chunked—and more on whether expertise still matters.
