Google's Spam Team Is Hunting Scaled AI Content. (But Using AI Isn’t the Trigger)
Google confirmed exactly what it's penalizing when it comes to AI-content — and named four sites that found out the hard way.
This week, Google's VP confirmed the spam team is still actively hunting scaled content. Four documented case studies show exactly what got hit — and it's not what most people think.
One of my clients recently brought up that they had seen a thread online about Google’s spam team going after AI content.
The fear? The client had been using ChatGPT to help draft close to a third of her blogs this year. And now she was asking me in a panic if she should delete it all or face the wrath of Google’s Spam Update.
I told her that she probably didn’t need to. But I was glad to see a bit of that concern that maybe using AI to scale content faster without a lot of intentionality was risky.
Her fear wasn’t dumb, either — because for the first time, we have receipts on what Google is and isn’t going after when it comes to AI on websites.
This week, Google’s VP confirmed the spam team is still actively hunting for sites that are obviously using AI to scale content to “hack” their way to the top of the SERP or trying to force-fit their site into AI answers.
Four sites were publicly documented as losing real rankings for the behavior.
And buried in Google’s own policy language is the actual answer to the question my client — and probably you — is asking. It’s not “did you use AI.” It’s something a lot more specific.
- Google's VP of Engineering confirmed it on the record — two words: “Yes it is.” The spam team is still actively hunting.
- Google's own policy: AI-generated content is only a violation when it's made to manipulate rankings and provides “little to no value … no matter how it's created.”
- SE Ranking: the August update caused an 82% jump in top-10 pages falling out of the top 100 — across all 20 industries tracked.
- Analyst Glenn Gabe documented four sites that got hit — losses of 200,000+, 14,000+, and ~25,000 queries, plus a 1.5-million-URL site.
- None of the four sites resembled a small business publishing AI-assisted content with human review.
- 74% of new web pages already contain AI-generated content — using AI was never the actual trigger.
What did Google confirm about AI-scaled content?
All of this started when SEO practitioner Gagan Ghotra asked on X whether Google’s spam team still had the resources to fight spam.
In particular, Ghotra wanted to know if Google was specifically calling out “scaled listicles” he’d seen ranking well. You know them — the “Top [thing] in [industry] for [year]" posts that are low-thought, high-copy, often full of slop.
Rajan Patel — Google’s VP of Engineering for Search — replied directly. Two words: “Yes it is.”
“Yes it is.”
— Rajan Patel, VP of Engineering, Google Search · responding to a question about spam-team resourcing, Aug 31, 2026
Patel’s response came at the end of a two-day rollout (Aug 18–21) that SE Ranking called the hardest spam update in recent memory — an 82% increase in top-10 pages falling out of the top 100 entirely, compared with a normal week.
A top-10 page was about 1.8x more likely to vanish from the top 100 during this update than during an average five-day stretch.
It hit all 20 industries SE Ranking tracks — real estate, fashion, healthcare, everything — with top-10 churn ranging from 74.6% to 85.6%.
But it wasn’t only losers. SE Ranking also found 12% more pages now breaking into the top 3 that hadn’t even cracked the top 20 before. So, some good news there!
But here’s what’s important to take away from all of this. Google didn’t say it was targeting AI content. It’s more nuanced than that.
more top-10 pages fell out of the top 100 during Google's August 2026 spam update
SE Ranking data via Search Engine Land — a top-10 page was about 1.8x more likely to disappear entirely, across all 20 industries tracked.
Google's rule, in Google's own words
Google updated its spam policy page three days before this update even shipped. Here’s the definition it landed on for what it calls “scaled content abuse”:
“Scaled content abuse is when many pages are generated for the primary purpose of manipulating search rankings and not helping users … no matter how it's created.”
— Google Search Central, spam policies for Google web search, last updated Aug 28, 2026
That last part stuck out to me: no matter how it’s created.
Google lists generative AI as only one way to do this badly. It’s right there in equal measure alongside old tactics like scraping content, spinning synonyms, and stitching pages together from other sites.
So, so far, the rule is not “Did you use AI?!”
The nuance here is “Did you make a lot of pages that don’t help anyone, using whatever tool was fastest?”
Four sites that got hit this month (and what they had in common)
SEO analyst Glenn Gabe published four case studies from this exact update on Aug 31. He blinds the sites — no names — but the pattern across all four is the whole point.
Four Sites, One Pattern
200,000+ queries, gone completely
An “ultra-YMYL” site (health, finance, safety) lost rankings for over 200,000 queries. Massive programmatic scaling with AI-generated content mixed in — a “two-headed monster,” per the analyst who found it.
14,000+ queries, 97% AI-detected filler
A fully automated Amazon affiliate site lost 14,000+ queries. Bottom-of-page filler text came back 97%+ probability AI-generated. Some categories had no product images at all.
1.5 million URLs, 85% low-value
A site with 1.5M+ indexed URLs — about 85% of them thin, programmatic pages. It had a legitimate human-run section too. Got hit site-wide anyway.
250,000 URLs, redirected to riskier sites
A site in a “risky niche” that funneled users to even riskier sites on click. Lost close to 25,000 queries. The analyst was surprised it hadn't been deindexed outright.
A site in an “ultra-YMYL” niche (health, finance, safety.
This is the category Google scrutinizes hardest) lost rankings for over 200,000 queries.
Completely, not just a drop. Gabe found “massive scaling via programmatic content” with chunks of AI-generated text mixed in — his words, “a two-headed monster.”
A fully automated Amazon affiliate site lost over 14,000 queries.
The filler text under its product listings came back 97%+ probability AI-generated.
Some categories didn’t even have product images. Gabe: “If anyone can set up a site programmatically like this with almost no human involvement at all, what value is there?”
A site with over 1.5 million indexed URLs, roughly 85% of them programmatic pages.
Gabe described them as having “no value-add … just a photo and some information you can find anywhere across the web.” It had some legitimate human-run sections too — didn’t matter, the whole domain still got hit.
A roughly 250,000-URL site in a “risky niche” that redirected users to even riskier sites when they clicked through.
This one lost close to 25,000 queries. Gabe said he was surprised it hadn’t been deindexed outright.
None of them is a small business with a dozen AI-assisted blog posts and a person reading them before they go live. Every one of them is a machine running at a scale no single person could review, built to fill space rather than answer a question.
The fair read
None of Google's four confirmed casualties this month were a small business owner using ChatGPT with a human editor reading the result. They were machines running with nobody home. If you're checking your AI-assisted drafts before they publish, this update was never built for you.
The line that matters
I wrote about this exact judgment gap back in July, before any of this was confirmed: 74% of new web pages already contain AI-generated content.
If using AI were the trigger, most of the web would already be gone.
But it isn’t the trigger. Volume without a human checking it is the trigger. It’s all about the content itself. Is it good? Helpful? Or just slop?
That’s good news if you’re using AI to draft, then reading it, fact-checking it, and deciding what’s worth publishing.
Now, I wouldn’t say that’s the best way to do this, but I hold no grudge against those who are using AI to their advantage here — particularly if you don’t love writing!
That’s exactly the layer none of Gabe’s four sites had. It’s bad news only if you’ve been treating AI as a way to skip that step entirely.
What to check on your own site
Does a real person read this before it publishes — not skim, read?
Does each piece say something a reader couldn’t get from five other AI-drafted pages on the same topic?
Is there an actual example, opinion, or number in it that’s yours — not just structure filled in?
If you strip your logo off the page, would a stranger know it’s you — or could it belong to anyone?
If you can answer yes to the first three and the fourth one bothers you, you’re not who this spam update was built for. Fix the fourth one anyway — it’s costing you more than rankings.
Do you actually know which of your pages would survive this test?
Most people react to news like this two ways — panic and delete a year of content, or ignore it and hope they're fine. Neither one is a strategy.
I'll tell you exactly where your content stands against what Google is actually enforcing — not the version of the rule everyone's scared of.
Fiverr Pro vetted · 4.9 stars · 1,600+ client reviews
FAQ
Will Google penalize my blog if I use AI to help write it?
Not for using AI. Google's own policy defines the violation as content made “to manipulate search rankings” and providing “little to no value … no matter how it's created” — generative AI is listed as one method among several, not a standalone violation.
What got hit in Google's August 2026 spam update?
SEO analyst Glenn Gabe documented four case studies — sites ranging from about 14,000 to over 200,000 lost queries — all involving massive, largely unreviewed content scale. None resembled a small business publishing AI-assisted posts with editorial review.
Did Google confirm it's still actively enforcing this?
Yes. On Aug 31, 2026, Google VP of Engineering Rajan Patel responded directly to a question about spam-team resourcing: “Yes it is.”
Does this affect AI Overviews and ChatGPT, or just classic Google rankings?
Per Gabe's reporting, spam-update impact carries through to AI Overviews and AI Mode, and can affect ChatGPT results when it grounds answers via Google Search.
Can a site recover after getting hit?
Yes, but it takes months of sustained cleanup after the underlying issue is fixed — not a quick fix. Gabe cited a real recovery example that took five months.
How This Was Made
Every claim above was checked directly against primary sources — Google's own policy documentation, Search Engine Roundtable, Search Engine Land, and Glenn Gabe's case studies — before publishing. Researched, directed, and signed off by Brad Bartlett — the name on this post is the name accountable for what's in it.
Written by
Brad Bartlett
Brad is a copywriter and content strategist who helps creators, brands, and organizations build content that's actually worth reading — and built to be found. He specializes in conversion-focused copy, brand voice, and SEO and AI search optimization, with a straightforward philosophy: great content has to be authentic before it can perform. He works comfortably across the AI content space, helping clients use the tools without losing the voice. Fiverr Pro vetted, 4.9 stars out of 5 across 1,600+ clients.