What's actually going on
Here's how it unfolds. Every time.
AI tools make publishing fast and cheap. You produce more content than ever before. Rankings start moving. It feels like the strategy is finally working. This is the most dangerous moment — because it feels like proof.
Growth slows. You publish more. The numbers stop responding the way they did at the start. You assume it needs more time — or more content. Neither is the real answer.
Google rolls out a core update. Sites doing real, structured work move up. Sites built on volume without substance drop — fast, silently, and without an explanation in any error report.
You look for what went wrong. Audit tools show nothing obviously broken. Rankings are down but there's no clear cause. You try more content. More links. Nothing moves. The industry has already moved on.
This is where most people end up when they find this page. There's still time. But the gap between where you are and where the recovering sites are gets wider every month you wait.
Why this happens
It's not that Google hates AI. It's what AI produces.
Google has been clear: AI-generated content isn't automatically penalised. What gets penalised is content that exists to fill space rather than serve a reader.
The problem is that AI tools, used without careful strategy, are very good at producing exactly that. Content that looks complete. Reads smoothly. Says the right keywords. And tells the reader almost nothing they couldn't have guessed without reading it.
Google's systems have been trained to distinguish between content that genuinely helps someone and content that performs helpfulness. That distinction is getting more accurate with every update.
Volume was never the goal. Usefulness was. AI made volume easy. It made usefulness no easier at all.Thin expertise
Content that covers a topic without demonstrating actual knowledge of it. Correct but shallow. Google's systems are increasingly able to tell the difference.
Recycled answers
AI trained on existing content tends to summarise what's already out there. Content that says what everyone else said — just slightly differently — adds nothing Google wants to rank.
Wrong intent match
AI content optimised for keywords rather than the specific stage of decision the reader is at. High traffic potential. Low conversion. Google notices when users leave unsatisfied.
No earned authority
Publishing volume builds content. It doesn't build trust. The off-page signals that tell Google your site deserves to rank — links, mentions, industry presence — don't appear because you published more.
Who this affects
You don't have to be using AI tools yourself.
The AI content problem affects businesses in more ways than the obvious one.
01
You're using AI tools directly
Publishing AI-generated content at volume. Rankings may be up now. The question isn't whether an update will come — it's whether your site will be ready when it does.
02
Your agency is using AI tools
You're paying for content and SEO work. You don't know how it's being produced. The risk is the same — you just don't have visibility into it.
03
Your competitors are using AI
Even if you aren't, this affects you. When an update rewards real content and penalises volume content, the rankings your competitors are holding artificially could open up — if your site is ready to take them.
04
You already got hit
Rankings dropped after an update. Traffic fell. The audit tools aren't showing a clear reason. AI content footprint is one of the first things I check in this situation.
What actually fixes it
The answer isn't to stop using AI. It's to stop letting AI lead.
AI is a production tool. It can draft, outline, format, and scale. What it can't do is replace the strategic thinking that tells you what to say, who to say it to, and why it matters to that specific reader at that specific moment.
The sites that weather algorithm updates are the ones where AI is a tool in the process — not the process itself.
Getting there requires going back to the foundation. Not deleting everything and starting over. Understanding what the market actually rewards, which content genuinely serves the reader, and how to rebuild authority signals that hold up under scrutiny.
I've worked with businesses at every stage of this — before the problem starts, mid-campaign when the plateau hits, and after the drop when the damage needs diagnosing. The earlier I get in, the more there is to save. But even after a drop there's usually a clear path. It's just longer.How I approach this specifically
A diagnosis before a prescription.
The first thing I look at is the competitive landscape — who's holding rankings in your space and what their content actually does for a reader versus what yours does. That comparison usually shows the gap clearly. Once the gap is visible, the fix isn't random. It's targeted. The content that's genuinely useful gets kept and strengthened. The volume content gets evaluated for whether it can be improved or whether it's doing more harm than good.
On the AI arms race
Some people's answer to Google's AI detection getting smarter is to use better AI tools that are harder to detect. That's a race you can't win long-term. Google's systems improve continuously. The only durable answer is content that genuinely serves the reader — which was always the goal, long before AI tools existed.
The honest part
Some sites are too far gone to recover quickly.
I won't pretend every situation has a fast fix.
Sites that have published heavily at volume for years, built their entire authority on that content, and compounded the problem with low-quality links pointing to it — recovery is a long project. Not impossible. But measured in months, not weeks.
This is why the timing matters. A site that addresses this before a major update hits is cleaning up. A site that addresses it after is rebuilding. One is maintenance. The other is reconstruction.
That said — I've seen sites recover from worse. The question isn't whether it's possible. It's whether you start now or wait until the gap is wider.
The businesses that are positioned well after the next update started their cleanup before it. Not the day it happened.