Google is banging its fist on the table: content generated at scale, regardless of the method, is in its sights. Between illusions of quality and the quest for traffic, a clarification was necessary. Do you publish en masse? Take two minutes. Really.
Key points:
- Google doesn’t discriminate: whether it’s human-generated or generated, mass-produced content is problematic.
- Content that is unoriginal or lacks real effort is now systematically ranked at the bottom of the quality scale.
- Using AI can make sense… if it serves to enrich an existing database, and not to produce generic content.
- Creating content solely to capture SEO traffic is a negative signal to Google.
A feeling of déjà vu?
If you thought AI content generation was going to be a thing of the past because Google said it didn’t care how it was produced, the company is telling you it’s not.
In a recent exchange, Danny Sullivan (Google Search Liaison) added another layer to the subject. It’s not the AI that’s problematic. It’s the intent behind it. And the scale. And the lack of added value. In short, you get the picture.
For Google, the problem isn’t new. It’s an old SEO habit: mass-produced content. Back then, it was done by hand. Today, it’s automated. The root of the problem remains the same: too much content with no substance.
What the Quality Raters Guidelines Really Say
A new look at quality
The famous Quality Rater Guidelines have evolved. AI-generated content is now directly listed in the “Lowest Quality” category if:
- Content is paraphrased, copied, or automatically generated without noticeable effort
- It offers no added value or originality.
- It’s just there to fill pages, without thinking about the user.
You don’t need an algorithm to sniff out bland content. Even quality raters are trained to do that.
No matter the tool
Sullivan insists: whether it’s done by hand, via AI, or with a home-grown script, the problem remains the same . If the goal is simply to rank, Google will sense it. And act accordingly.
And let’s be honest. Remember the sites stuffed with directory links in 2008? These kinds of practices are making a comeback, but with a layer of prompting.
But then, what do we do with AI?
Let’s not throw the baby out with the bathwater. Google doesn’t hate AI. What it wants to avoid is a flood of soulless content.
Sullivan gives an example: On Amazon, automated customer review summaries are useful. They synthesize authentic feedback. And they provide clarity. That’s real added value .
A question of intention
Before you click “Publish 500 pages,” ask yourself, “Am I doing this for my audience… or for Search Console?”
If you don’t think a real reader would gain anything from stumbling upon this page outside of Google, maybe this page doesn’t belong there. It’s harsh, but that’s the message.
Originality or illusion?
A striking example: those who praise AI-generated content quickly change their tune… when it comes to a subject they understand. You can immediately tell when it rings hollow.
Why? Because AI doesn’t think. It strings together probable words. And that can give the illusion of relevance. Until you dig a little deeper.
- Content that skims over without going into depth
- Smooth, but empty sentences
- Generic answers to overly specific questions
And just because it “reads well” doesn’t mean it’s “worth anything.”
Before generating: 2 tests to do
Sullivan suggests two questions to ask yourself. Simple. But effective:
- Why am I creating this content? To help, or to attract traffic?
- What am I bringing that’s new? Am I just rehashing what already exists?
If the answers make you uncomfortable…you probably have your answer.
We’re not telling you to ban AI. We’re telling you not to use it like a photocopier. Write for people, not robots. It seems obvious. But in the rush to rank and calculate KPIs, we forget it too quickly.
And if you feel tempted to publish 300 “SEO-friendly” pages at once… take a deep breath. Sit down. And ask yourself if anyone would actually want to read them.
