AI Content vs Human-Written Content: I’m Testing Both

There’s another problem too: how do you actually know?

You can pay a writer to produce an article, but that doesn’t necessarily mean AI wasn’t involved. Unless you know how the content was produced, it’s becoming increasingly difficult to be completely sure.

Personally, I haven’t been able to make up my mind.

I’ve tested different approaches in the past, but I’ve never been able to identify a clear pattern. I’ve seen AI-generated content perform extremely well and I’ve seen it do absolutely nothing.

But I’ve also seen exactly the same thing with content written by professional writers – and content I’ve written myself.

So instead of having another opinion about it, I decided to run my own experiment.

Within two weeks of each other, I launched two new iGaming websites.

One is 100% human-written.

The other is 100% AI-generated.

And I’m going to document what happens.

Project 1: 100% human-written

For the first project, I went premium.

I hired writers I already know and have worked with before, and paid just under €4,000 for the initial content for a completely new domain.

The content is 100% human-written. No AI was used in the writing process, and this is something I’ve been able to verify.

This isn’t an expired domain or something with an existing history or backlink profile. It’s what we’d normally call a hand-registered domain.

The site launched on September 4, 2026.

I’ve deliberately paid for writers I trust rather than trying to find the cheapest possible option.

In other words, if human-written content is supposed to have an advantage, I’m trying to give it a fair chance.

Project 2: 100% AI-generated

The second project takes the opposite approach.

Everything is AI-generated.

Reviews, news articles, landing pages and supporting content – all of it.

My human involvement on the content side is deliberately limited. I did some basic keyword research, decided what content I wanted and published it myself.

That’s basically it.

The question I’m trying to answer is simple:

Can you build and rank an iGaming website using little more than AI, WordPress and link building?

And if you can, how does it compare with a site where I’ve spent almost €4,000 on human writers?

That’s what I want to find out.

This isn’t a perfect experiment

There are some obvious weaknesses in the setup, and I think they’re worth pointing out from the beginning.

The two websites aren’t targeting the exact same country.

I’ve deliberately chosen two relatively small markets that are similar in terms of language and overall market size. I needed markets where a brand-new site actually has a realistic chance of getting somewhere within a reasonable amount of time.

But that introduces a variable I can’t completely control.

Google may understand one language better than the other. AI-generated content may also perform differently depending on the language and how well current models handle it.

Competition won’t be identical either.

So this isn’t a scientific A/B test where every variable except the content is perfectly controlled.

I’m not going to pretend that it is.

What I can do is give the two projects as similar a starting point as possible and document exactly what happens.

And to be fair, there needs to be some chance that the AI project actually succeeds. Otherwise, this wouldn’t be a particularly interesting experiment.

What do I think will happen?

I genuinely don’t know.

That’s also the reason I’m doing this.

Part of me hopes the human-written site wins.

For the future of Google – and probably for the future of SEO – I’d like to think there is still some value in paying someone who understands a subject to sit down and actually write about it.

But there’s another side to this.

If the 100% AI-generated site performs just as well, that makes life significantly easier – and cheaper – for anyone running affiliate websites.

It’s difficult to ignore the economics.

One project starts with almost €4,000 invested in content.

The other can produce content at a fraction of that cost.

If their performance ends up being relatively similar, that’s a pretty important result in itself.

How I’m going to document the results

I’m going to publish an update on both websites every month.

I’ll be tracking indexing, organic traffic, rankings and FTDs. More importantly, I want to see what happens to both sites when Google updates roll through.

I’m also going to try to keep the link-building side as comparable as realistically possible.

The idea isn’t to build great links to one website and cheap links to the other and then declare a winner. Both projects should get roughly the same opportunities from an off-page perspective, and I’ll document what I’m doing along the way.

At least initially, however, I’m going to keep some details private.

I won’t reveal the domains or the exact websites I’m building links from. The markets, domains and link sources will therefore be anonymised in the reports.

That’s partly to protect the experiment itself. Once the domains are public, there’s always the possibility of outside factors influencing the results.

The first update comes in October

The first proper update will come in the second week of October.

By then, both websites will have had some time to settle and I should hopefully have the first useful data.

I’m particularly interested in indexing at that stage.

Does Google discover and index the content at roughly the same rate? Does one site get more pages indexed? Does one start picking up impressions and rankings faster?

Or will there be no meaningful difference at all?

I have no idea yet.

And that’s really the point of the experiment.

Rather than arguing about whether AI content works, I’m going to build both and see what Google does with them.

Experiment updates

For the record, I’ll keep this article updated with links to each new update from the experiment below. That way, you can follow both sites from launch and see how the results develop over time.

October 2026: Coming soon