The scaled AI content that gives your traffic a spike today is exactly what Google’s spam updates penalise. That’s not scaremongering, it’s policy in black and white. And it’s a temptation many brands fall for right now, because it has become so easy to have hundreds of pages written in a short time. It works too, for a while. That is precisely where the danger lies.
What Google actually penalises
Google calls it “scaled content abuse” and defines it plainly: many pages created mainly to manipulate rankings rather than to help users. In their own spam policy , the use of generative AI to mass-produce pages without adding value is named explicitly as an example.
Note the wording: it is not about AI as such. It is about scaling without value. In the combined core and spam update of March 2024, Google’s own goal was to cut the amount of low-quality content by 45%. That is not a tweak at the margin, that is halving the noise.
We don’t build pages, we build logic.
Giacomo Perticara, founder and SEO strategist at GRP Digital
The spike, then the fall
The pattern is well documented by now. SEO researcher Lily Ray tracked more than 220 sites that worked with AI content tools. 54% lost more than 30% of their peak traffic, 39% lost more than half, and 22% lost more than 75%. The course is almost always the same: rapid page growth, a traffic peak three to six months later, and then a steep fall within a year. The title of her analysis sums it up: “It Works Until It Doesn’t.”
Why your AI citations follow the fall
Now the part people often miss. When your organic visibility collapses, you take your AI mentions down into the abyss with it. That is not theory. Lily Ray tracked eleven sections of large brands that lost organic traffic, between -5.7% and -53.1%. Across all AI platforms combined, their citations dropped by 22.5% on average. Every section that lost organic ground also lost AI citations. Her conclusion: the fastest way to lose visibility in AI is to lose it in Google first.
The logic behind it we know from the rest of this story: AI models draw on the same search indexes. Drop out of the organic results and you disappear from the source the answer is built from. We explained that mechanism earlier in why AI makes SEO more important, not less.
The honest nuance
This link is strongest with Google’s own AI and with Perplexity, which work live from the index. With ChatGPT the connection to your Google ranking is looser, because it leans partly on its own knowledge and its own retrieval. But the common thread holds, and it is called authority. Domains with a lot of trust get thousands of AI mentions, while weak domains are barely cited. Trust and authority determine both your classic ranking and your AI visibility, and that is exactly what scaled, thin content fails to build (E-E-A-T).
Don’t build for today’s loopholes
The temptation is understandable. A spike is visible, fast and easy to sell. But you are then optimising for the temporary weakness of a system, and behind your site sits a real business with real revenue. Building on something that gets penalised again in a few months is not a strategy, it is a loan at a high interest rate.
At GRP Digital we deliberately choose the other side. Not because the slow road sounds nobler, but because it is the only one that keeps standing. Content that genuinely adds something, built on authority you earn rather than fake, performs in Google and in AI answers. That is also why we argue for your own organic growth instead of rented visibility. And when the next update comes around, you have nothing to fear.
Frequently asked questions
Does this mean I can’t use AI for content at all? You can. Google does not penalise AI, but the mass production of pages without value. AI as a tool for content that genuinely adds something is fine; AI as a machine to churn out thin pages is not.
Why should I care about Google if I want to score in AI? Because AI answers draw on the same search indexes. Lose your organic visibility and your AI mentions drop with it.
How long before scaled content gets penalised? Often three to six months of growth, followed by a fall within a year around a core or spam update. The exact moment can’t be predicted, the direction can.
Sources: Google Search Central (spam policy, scaled content abuse), Amsive / Lily Ray (March 2024 update) and Lily Ray, “It Works Until It Doesn’t” en on AI citations and organic visibility.
