The Truth About AI-Generated Content and Google Rankings

This morning, over on LinkedIn, I posted an April Fools joke.

It was about Google and LinkedIn banning AI-generated content.

The truth is more shocking than the fantasy.

Not because AI content is being banned, but because it plainly is not.

Ahrefs analysed 600,000 pages and found that 86.5% of top-ranking pages contained some level of AI-generated or AI-assisted content. More importantly, the correlation between AI-content percentage and ranking position was effectively zero. In plain English, Google does not appear to reward or punish content just because AI helped create it. Ahrefs explains the findings here.

That should put one tired argument to bed.

AI Content Is Not the Problem. Bad Content Is.

Google’s own guidance says generative AI can be useful for research and structuring original content. What Google warns against is using it to generate lots of pages without adding value for users. That is the line that matters. Not whether the first draft started with a keyboard or a prompt box. Google says so in its guidance on using generative AI content.

That is why I keep saying AI works best when it is powered by a human.

And the Ahrefs data backs that up. Ahrefs reports that 87% of marketers use AI to help create content, 97% edit AI content before publishing, and websites using AI content grew faster than those that did not. That does not mean AI is magic. It means smart people are using it to move faster on research, structure, first drafts, updates, and rewrites, then bringing human judgement to the finished piece. Their research summary is here.

I have said before that an idiot using AI is just a more powerful idiot.

I still believe that.

But someone who takes the time to learn the tools, question the output, and improve the final result can become a very powerful operator.

That is the real divide now.

Not AI users versus non-users.

Capable operators versus sloppy ones.

Ahrefs also found that most new web content now carries some AI footprint. So this is not a fringe habit being practised in secret. It is rapidly becoming normal workflow. You can read that Ahrefs research here.

The Em Dash Is Not the Enemy

Which brings me to the em dash. And the rocket emoji. And all the other little things people wave around as if they are holy water against machine-written text.

No, the em dash is not the enemy.

The rocket emoji is not the enemy.

Weak thinking is the enemy. Lazy writing is the enemy. Publishing bland, unchecked pages at scale is the enemy.

Some copywriters talk as if they can sniff out an AI-written post on sight. Sometimes they probably can. Sometimes they are just spotting bad writing and giving AI all the blame. Those are not always the same thing.

To be fair, I understand why some writers are nervous. AI is a genuine commercial threat to anyone whose clients are now wondering how much of the writing process they can do for themselves. That threat is real, and I respect it.

I also suspect plenty of people criticising AI in public are quietly using it in private to work faster and tighten their process. That is not hypocrisy so much as survival.

Is AI Content Really Unoriginal?

Another copywriter I respect has said that AI-generated content is unoriginal because it is based on what is already out there. There is some truth in that, but not as much as people think. Most writing starts there. Very little content is created in a vacuum. Writers research, absorb, interpret, and then shape something useful for a specific audience.

When I was in school, I used encyclopaedias to research reports. I did not invent the underlying facts myself. I gathered material, made sense of it, and wrote the report. AI is not exactly the same, but it is not miles away either. It is a modern research and drafting tool. The difference is that it moves faster, sounds more fluent, and can make a mess more quickly if you use it badly.

And that really reinforces the point. AI is not the issue. The human behind it is what makes it work. Whether that human is me as an SEO strategist, a copywriter, or a small business owner, the same rule applies. The tool can help, but judgement is still doing the heavy lifting.

So the real question is not whether the information existed before. Of course it did. The real question is whether you add judgement, accuracy, clarity, and usefulness to the final piece. That is where the human still matters.

Rebecca’s Fork Analogy Nails It

My friend Rebecca Harrison, a copywriter, gets this. She has a good analogy. When new tools arrive, people mock them, fear them, then quietly adopt them once the panic fades. She points to the humble fork. At one stage it was treated with suspicion. Later it became normal. Then it became refined. Same tool. Different public mood.

AI is going the same way.

ChatGPT, Claude, Gemini, Jasper and the rest are tools. Used well, they can be an essential part of modern work. Used badly, they can do real damage. A fork is part of daily life when you use it properly. Poke yourself in the eye with it, though, and you will form a very different opinion. The tool was never the real issue. The person using it was.

Why Would Google Fight Its Own Tools?

There is another reason the old “Google hates AI content” line has always sounded a bit suspect.

Google is not exactly behaving like a company at war with AI-assisted creation. Its own Search documentation explains how generative AI can be used responsibly. It would be odd for Google to spell out responsible uses of AI content, then punish people simply for using it properly. Again, Google’s own documentation makes that clear.

That does not mean Google gives AI content a free pass.

It means Google is still judging output, not throwing a tantrum over the tool.

The same pattern shows up in AI Overviews. Ahrefs found that 91.4% of pages cited in AI Overviews contain some amount of AI-generated content. So if anyone is still clinging to the idea that AI-assisted pages are somehow locked out of modern search visibility, the evidence says otherwise. Ahrefs covers that here.

The Real Risk Is Careless Use

There is a warning label here, though.

Ahrefs also found that human content was slightly more likely to be negatively impacted by a Google update than AI content. At the same time, AI does make it easier to produce scaled rubbish quickly, which is exactly the sort of thing Google warns against. So the real risk is not AI itself. The risk is careless use of it. Google’s guidance is still focused on quality and value.

So no, this is not an argument for filling your site with machine-made sludge and hoping for the best.

It is an argument for using better tools with proper standards.

The Same Logic Applies to Images

And yes, I think the same principle applies to images.

From time to time, people claim that audiences do not like AI-generated visuals. Maybe in some cases they do not. But I have yet to see convincing evidence that people consistently reject an image simply because AI played a part in making it.

Where exactly is the line now?

If you use Photoshop, Canva, generative fill, background cleanup, smart resizing, or other assisted tools, is that somehow pure and noble, but using an image model is not?

That feels less like a principle and more like people choosing their favourite shovel and sneering at the newer one.

Again, the better question is not, “Was AI involved?”

The better question is, “Is the result any good, is it honest, and does it do the job?”

That is the standard that matters in writing, design, and search.

No Fooling

So here is the truth behind the joke I posted this morning.

AI-generated content is not being banned.

It is already embedded in the modern web, in search results, in marketing workflows, and in the tools used by the people creating content every day. Ahrefs’ data makes that plain, and Google’s own guidance points in the same direction.

The winners will not be the people who avoid AI out of fear.

They will be the people who learn how to use it without lowering their standards.

That is the game now.

And if your entire anti-AI argument still rests on an em dash, you may have bigger problems than punctuation.

If you want a human-guided approach to AI-powered SEO, have a look at The Norman Advantage.