Most businesses adopting AI-generated content are not cutting corners. They’re just trying to move faster. The problem is that the types of content that AI makes easiest to produce at scale are the ones Google has become most effective at penalising. That and the lag between publishing and performance collapse can run to 12 months or more, meaning the damage is often invisible until it is already widespread.
The issue is not AI itself. It is a small set of formulaic content patterns that AI can cheaply and quickly replicate, and that Google's quality systems are specifically designed to identify. Volume-first content is not a neutral choice. It carries documented risk, and understanding where that risk sits is the starting point for doing this differently.
This post draws on research covering more than 820,000 data points and Google's own content guidance to set out where the risk sits, and what a more durable approach looks like.
The Rank & Tank Traffic Spike of Mount AI
Scaling content with AI is not a low-risk strategy for organic search. According to Search Engine Land (SEL), 73% of B2B websites saw enormous traffic losses between 2024 and 2025, with an average year-on-year decline of 34%. Sites publishing primarily informational content have been hit hardest, with some sectors reporting organic traffic declines of 15% to 64% since AI Overviews launched [1].
SEO consultant Lily Ray's analysis of more than 220 monitored websites, published in Search Engine Journal (SEJ), found that 54% of sites that had scaled AI content lost 30% or more of their peak organic traffic. 39% lost more than half. SEO professionals refer to this pattern as "Mount AI", AKA, a rapid growth in indexed pages, a traffic peak, then a steep decline that frequently drops below the original baseline [2].
For businesses that have already seen organic traffic flatten or fall, understanding how SEO and paid search work together is often the first practical step toward recovery.
The 8 Content Patterns Google Is Already Penalising
Lily Ray's research also identified eight recurring content patterns associated with organic traffic declines. Most affected sites were running at least three or four simultaneously. The most aggressive were running all eight.
Comparison Pages Published at Scale
Product A vs Product B articles published across every conceivable head-to-head matchup in a category. The template is formulaic and easily replicated by any competitor running the same prompt, which is precisely why Google has grown effective at identifying it.
"What is X" Glossary Pages Scaled for AI Citation
Single-term pages designed to be cited by AI engines, often scaled programmatically across multiple languages from one source template. At scale, they create a low-quality content footprint that weighs down an entire domain.
"Best X for Y" Listicles & The Affiliate-Era Template
The most familiar template in the AI content playbook, with roots in the affiliate-content era. It’s been observed across both broad-category and narrow-niche variants. The pattern is now so widely replicated that it functions as a signal in itself.
The Self-Promotional Listicle & The Update that Followed
A variant of the above, where the publisher ranks itself as the top option in a category it competes in. These pages rarely include evidence of genuine testing, which Google's review content guidance explicitly requires. A wave of sites has been hit by an unconfirmed Google update from 20 January 2026, with traffic drops of 40% to 95% across the January to April 2026 window.
Competitor-vs-Alternatives Pages Are Built to Intercept, Not Inform
Dedicated pages targeting every named rival in a category at scale. For example, let’s say the majority of a site's highest-traffic pages were built around individual competitor brand names. The content exists to intercept searches, not to help readers make informed decisions.
Programmatic Location & Language Scaling at Volume
One template multiplied across every town, city, or language a search engine will index, with near-zero unique content per page. This approach has triggered algorithmic action for over a decade. AI has made it faster to produce, and Google has accelerated its response accordingly.
FAQ Farms Scaled for Schema, Not for Readers
One URL per question, structured for AI extraction, with schema at the bottom. At scale, this creates erroneous low-quality content baggage across a domain. Google's decision to deprecate FAQ Rich Results appears connected to the surge in FAQ schema aimed at earning AI search citations.
Off-Topic Content at Scale & The Damage to Topical Authority
High-volume articles with no genuine connection to the publishing business. Off-topic content at scale damages topical authority, the signal that tells Google your site genuinely covers a subject in depth. That damage extends beyond the off-topic pages themselves to the broader domain.
What Google Actually Rewards (& Why It Requires More Than a Prompt)
Most AI content discussions miss a key distinction. Google does not penalise AI-generated content as a category. Ahrefs analysed 600,000 pages across the top 20 ranking positions for 100,000 keywords and found a correlation of 0.011 between AI content percentage and ranking position [3].
For context, that is so close to zero it is basically waving at zero from across the street. Yet what Google penalises has only recently become much clearer.
Its own guidance identifies three types of content its systems are built to surface:
- Content that draws on genuine first-hand experience, with real decisions and real context.
- Material that adds something a reader could not find by running the same search themselves.
- Copy that exists because it serves the reader, not because it targets a ranking or an AI citation.
The opposite of all three is commodity content, and Google's own example of the problem is a post titled "7 Tips for First-Time Homebuyers." AI tools can support the creation of better content, but the judgement, expertise, and editorial oversight have to come from somewhere [4].
That is where working with the right team on your content strategy and copywriting makes a practical difference, and where bespoke AI software built around your actual workflows adds value that off-the-shelf tools cannot replicate.
The Businesses Still Ranking Have One Thing in Common
Without a clear view of what is driving your content performance, it is hard to know whether your results will hold or collapse when Google's quality systems catch up. The businesses holding their rankings are the ones whose content reflects genuine expertise. We’re talking about material that couldn't have been generated by the same AI prompt a competitor is already using. That is the standard this article has been pointing towards, and it is achievable with the right strategy.
b4b is a full-service digital marketing agency based in Poole, working with clients across the UK and internationally on AI, software, and websites. With a 4.9 Google rating from 89 reviews and hundreds of completed projects, our team understands how Google's quality systems work in practice and produces content that performs over time, not just in the first few months.
Call 01202 684400 or get in touch to talk through your content strategy and get ranking for what matters in your sector.
External Sources
[4] Google Search Central, Optimizing your website for generative AI features on Google Search (2026)
