A fab study that appears to show that AI content doesn’t hurt your Google rankings is doing the rounds on social media. Cue the stampede of businesses rushing to pump out hundreds of AI-generated articles faster than you can say “content at scale.”

Hold your bloody horses.

Yes, the Ahrefs study of 600,000 pages is brilliant. Yes, it shows that 81.9% of top-ranking pages contain some AI content. But here’s what everyone’s conveniently ignoring: there’s a world of difference between using AI as a tool and using AI to create masses of content just because you can.

What the study shows

The Ahrefs research analysed 600,000 pages from 100,000 random keywords, looking at the top 20 ranking URLs. Here’s what they found:

4.6% of pages were pure AI content
13.5% were pure human content
81.9% were a mix of both

The correlation between AI content percentage and search ranking was 0.011 – effectively zero. This suggests Google neither rewards nor punishes AI content based on how it’s created.

(BUT – let’s not forget that we don’t know how accurate the AI checker Ahrefs used is, and there have been a lot of articles about how shit AI detectors are. It’s worth reading The Dark Side of AI Detectors: Why Accuracy Is Not Guaranteed and The Truth About AI Detectors – More Harm Than Good when you have a mo.)

Smart AI use vs lazy AI use

Think of AI assistance like using a calculator versus having someone else complete your entire maths homework. Using a calculator to check your work or speed up calculations is smart and widely accepted. But having someone else do your entire homework means you’re not learning, not adding value, and you’ll eventually get caught when the teacher asks you to explain your work.

What responsible AI use looks like

Smart businesses are using AI to enhance their expertise, not replace it:

An IT consultant using AI to improve the flow of their technical explanations
A marketing agency asking AI to help structure case studies based on their client results
An accountancy firm getting AI to create clearer explanations of complex tax changes their readers will understand
A business coach using AI to refine their methodology into more digestible content
A web design agency utilising AI to help explain technical concepts to non-technical clients

The key? These businesses are adding their knowledge, experience, and personality to AI-assisted content.

How not use AI for SEO

Meanwhile, others saw “AI content doesn’t hurt rankings” and interpreted it as “let’s create 200 articles about business coaching in every UK city”:

Publishing 50+ AI-generated articles per week without human review
Creating hundreds of location-based service pages (business coaching Leeds, business coaching Birmingham, etc.)
Using AI to rewrite competitor content with minimal changes
Generating entire websites of content in industries they know nothing about
Publishing AI content without fact-checking or adding any unique insights

This very definitely ISN’T the way to do things.

Why Google’s patience has limits

Google’s been clear: they care about quality, not how content is created. But there’s a difference between using AI to enhance your expertise and using it to manufacture fake expertise at scale.

As Glenn Gabe pointed out when the study was published: “Please do not read this post and run to publish AI-generated content at scale. There is nuance with this topic… There are many examples of sites that got destroyed via broad core updates or spam updates that went down the path of heavily using AI content to scale.”

Tweet from glenn gabe - it says: please do not read this post and run to publish ai-generated content at scale. There is nuance with this topic that i believe the post misses... There are many examples of sites that got destroyed via broad core updates or spam updates that went down the path of heavily using ai content to scale. Oh, and manual actions too. See the screenshot below of one example. Google has always explained it's about quality, and not that's it's just ai content, but using pure ai-generated content to scale heavily can be super dangerous. Also, the post misses the important point of google evaluating sites *overall* with broad core updates. So you can see positive gains until you tip the scales. Then look out. Beware. I have written and presented about site-level quality algorithms many, many times.

Sites that went heavy on AI content scaling have been getting demoted and even removed from SERPs completely in broad core updates and spam updates. Some have even received manual actions. Why? Because Google evaluates sites overall, not page by page.

You might see positive gains initially, but eventually you “tip the scales”. Then you’re properly f*cked.

The nuance everyone’s missing

When Google says AI content is fine, they’re not giving you permission to become a content farm. They’re saying that using AI tools responsibly won’t hurt you.

The study shows that most successful content uses some AI assistance – but the key word is “assistance.” The best-performing content still has human expertise, insights, and personality behind it.

What you should do

Use AI to make your expertise more accessible, not to fake expertise you don’t have:

Let AI help you explain complex concepts more clearly
Use it to improve grammar and flow of your insights
Get AI to suggest better headlines for your expert content
Have it help structure your knowledge into readable formats
Use it to create variations of your proven content for different platforms

But always add your own knowledge, experience, and perspective. If you can’t explain or defend what you’ve published, you shouldn’t have published it.

The Ahrefs study suggests AI assistance works. It doesn’t prove that mindless AI scaling works. There’s a massive difference, and if you ignore it, Google will eventually remind you in the worst possible way.

Sick of watching competitors pump out AI rubbish while you’re stuck wondering what’s safe to do? Let’s have a proper chat about using AI responsibly to grow your business without risking a Google penalty.