AI-Generated Content Loses Its Luster: A Cautionary Tale of Mount AI

I worked with a brand a few years ago that generated thousands of articles with AI automation.

I still have access to their GSC, so I took a look back recently to see how that turned out.

This is what happened to their Google organic clicks and impressions:

The fall in clicks and impressions coincides with Google’s June 2025 Core Update.

(I can’t say the core update caused the decline, nor can I prove the site’s use of AI-generated content was the reason it lost visibility. I’m simply noting the correlation.)

You might be wondering: “Why didn’t you advise them on another way, like creating helpful, reliable, people-first content?”

I did …

In our first meeting, I explicitly stated that the approach being taken to AI-generated content production was a long-term risk, despite the short-term windfall they were getting for organic traffic.

Our relationship didn’t last long, but they were good people, and I sincerely hope they took my other advice, which was to build up their other marketing channels (social, email, paid media, etc.) and not to rely solely on organic traffic.

But that one example isn’t isolated.

You don’t have to look far to find other cautionary tales.

I see them pretty much every day on X:

A short story: pic.twitter.com/J5ixA0wFjV

— Harpreet (@harpreetchatha_) August 13, 2026

This phenomenon is what a lot of SEO/GEO practitioners call “Mount AI”:

Mount AI describes a pattern where a site publishes large amounts of AI-generated content and then has strong organic growth (temporarily) followed by a precipitous decline in search visibility. Such sites may violate Google’s scaled content abuse policy, namely if the content was created to manipulate search rankings (rather than help users).

These Mount AI cautionary tales have brought up questions around the wisdom of AI-generated content, in general.

For example, I was recently looking at a subreddit I follow for AI search beginners, where I like to offer advice from my own experience.

Here’s a comment I left on a question about whether using AI-generated content was a risk for SEO:

In short, I don’t believe all AI-generated content is an inherent risk for SEO/GEO.

It’s content that lacks purpose and value that’s a risk, so-called “commodity content.”

As Google says in their guidance on AI-generated content:

“Google’s ranking systems aim to reward original, high-quality content that demonstrates qualities of what we call E-E-A-T: expertise, experience, authoritativeness, and trustworthiness. We share more about this in our How Search Works site.

Our focus on the quality of content, rather than how content is produced, is a useful guide that has helped us deliver reliable, high quality results to users for years.”

In other words, Google isn’t opposed to AI-generated content; it’s opposed to content designed to manipulate rankings (like spam content).

Do I use AI-generated content myself?

Surely.

I have a series in my blog called Hamsterdam Research, where I look at AI research papers to find insights for SEO/GEO. Many of the papers I examine are academic and hard for me to understand, so I rely on an AI assistant like ChatGPT to distill their information. Sometimes, I include ChatGPT’s summaries directly in the blog posts I create.

The reason my inclusion of ChatGPT-derived content into my Hamsterdam Research blog posts isn’t “commodity content” though is because I take steps to make the content “satisfying,” like by offering a unique perspective in my analysis of the research that (hopefully) contributes to the knowledge in my SEO/GEO niche.

And I’m not the only one using AI-generated content.

In 2025, Ahrefs found that 74% of new webpages include AI-generated content.

In another study done in 2025, Ahrefs found:

“According to our AI detector, Human-written content only made up 13.5% of the top 20 ranking pages. This means 86.5% of top-ranking pages contain some amount of AI-generated content.”

The takeaway from the Mount AI examples isn’t that AI-generated content is inherently bad for your SEO/GEO; it’s that commodity content at scale can put your traffic at risk.

Thus, every page you create should have a purpose and offer unique value, like this page does by including the case study of the brand I used to work with — that’s unique information based on my personal experience.

Don’t fall into the trap of mass content production with AI. Don’t be a Mount AI cautionary tale.

Instead, use AI if you must for your content, but just ensure you’re also incorporating experience or expertise to contribute unique value based on having a distinct purpose for creating the content.

One way to accomplish this is to create a RAG pipeline where the AI assistant or agent is pulling from original research or other high-value proprietary content you’ve produced. When used sensibly, this technique can help the content stay original and unique to you (and be valuable to your audience).

Until next time, enjoy the vibes:

Thanks for reading. Happy optimizing!

Related posts