There's a belief in SEO that has hardened into something close to religious conviction: that human-written content is inherently better than AI-written content, and that the act of a human typing the words is itself a mark of quality. People put 'written by humans' badges on their sites. Agencies charge premiums for '100% human-written' guarantees. And buyers nervously ask whether the content attached to their links was written by a person, as though that single fact settles the question of whether it's any good. It doesn't. The premise is wrong, and clinging to it leads you to pay for the wrong things and worry about the wrong risks.

To be clear up front: this is not an argument that AI content is great and you should mass-produce it. The opposite failure mode — flooding a site with unedited AI filler — is genuinely dangerous, and we'll get to exactly why. The point is narrower and more useful: who or what physically produced the words is not what determines whether content works. Quality, originality, and genuine usefulness determine that, and a human can fail those tests just as easily as a machine can. Origin is not quality. Confusing the two is the myth worth killing.

What The Data Actually Shows

Start with the evidence, because the 'human is always better' position tends to evaporate on contact with it. When Semrush studied AI versus human content at comparable quality, they found near-identical ranking performance: roughly 57% of AI articles and 58% of human articles appeared in Google's top ten. That's a rounding error, not a chasm. Comparable quality produced comparable results regardless of who wrote it.

Google's own position says the same thing, and has for years. Its guidelines focus on whether content is helpful, reliable, and created for people — not on whether a human or a machine produced it. Google has stated plainly that using AI is not against its guidelines; using automation to mass-produce unoriginal content to manipulate rankings is. The distinction is about quality and intent, not authorship. A search engine does not award points for the words having been typed by a human, and it never has.

If human origin guaranteed quality, you would never see a thin, useless, human-written page — and the web is drowning in them. Conversely, genuinely useful AI-assisted content ranks perfectly well. The thing the badge is supposed to certify simply isn't the thing that matters. What matters is closer to what the Helpful Content Update actually targeted — which was never about catching AI, but about demoting content that existed to capture search traffic rather than to help anyone, whoever or whatever produced it.

So Why Does AI Content Get Sites Penalised?

Here's where the nuance lives, because AI content genuinely does sink sites — just not for the reason most people think. The famous cautionary tale is the 'SEO heist': a site generated roughly 1,800 AI articles by copying a competitor's topic list, briefly pulled in nearly half a million visits a month, and then got hit with a manual penalty that took its traffic to almost nothing. People pointed at it and said 'see, AI content gets you killed.' That's the wrong lesson.

The site wasn't penalised because the content was AI-written. It was penalised because it was mass-produced, unoriginal, and created purely to capture traffic at scale — it added nothing that didn't already exist. Those exact same articles, written by a room full of humans copying a competitor's topics with the same goal, would have been just as worthless and just as liable to demotion. The AI made the bad strategy faster and cheaper to execute; it didn't invent the badness. The cause was the content's quality and intent, not its origin. This is the whole confusion in miniature: people blame the tool when the actual problem was producing undifferentiated junk on an industrial scale.

The Real Problem With Lazy AI Content: Convergence

There is a genuine, specific danger with carelessly generated AI content, and it's worth understanding precisely because it's so often described wrongly. The lazy version of the warning is 'AI produces duplicate content.' That's not quite right — ask the same AI the same question twice and you'll get slightly different wording each time, so you won't get byte-identical pages. The real problem is subtler and worse: convergence.

When thousands of people prompt the same AI tools with similar questions about the same topics, the outputs converge on the same structure, the same points, the same framing, the same safe middle-of-the-road answer. No single pair of articles is identical, but the whole genre collapses toward a generic average. Your AI article about 'the benefits of X' is not a duplicate of your competitor's AI article about 'the benefits of X' — it's just indistinguishable from it in every way that matters. And that raises a question you cannot answer: if your page says exactly what a user would get by asking the AI directly, why would they read your page, and why would Google index a thousandth near-identical version of the same answer?

This is the actual risk, and it's a quality-and-differentiation problem, not an authorship problem. It's also exactly why thin, generic content increasingly struggles to get indexed at all — the subject of why Google increasingly refuses to index thin content at all. A page that adds nothing beyond what already exists is a page Google has every reason to leave out. The convergence problem is real; 'a human didn't type it' is not the reason it's real.

What Actually Separates Content That Works

If origin isn't the dividing line, what is? The honest answer is the same set of things that separated good content from bad long before AI existed — they've just become more decisive now that average, comprehensive content is free to generate.

Genuine Experience

There's a useful distinction buried in Google's E-E-A-T framework: the first E is Experience, and it's the one thing a machine genuinely cannot fake. An AI can synthesise expertise — it can assemble what's known about a topic competently. What it cannot do is have actually used the product, visited the place, run the test, made the mistake, or lived the situation. Content carrying real first-hand experience — specific details, things that only someone who'd actually done it would know — reads as different, and increasingly ranks as different, because it contains information that doesn't exist anywhere else. This is true whether a human writes it unaided or uses AI as a drafting tool, as long as the genuine experience is actually in there.

Information Gain

The content that wins adds something to the conversation rather than restating it. A genuinely new data point, a contrarian but defensible angle, a specific example, an actual opinion — anything that means the page isn't just a competent re-articulation of the existing consensus. Average comprehensiveness is now the floor that anyone can reach with a prompt; the differentiator is contributing something beyond it. A human who only ever restates the obvious clears this bar no better than a machine does.

Being Genuinely Useful To A Real Reader

The final test is the oldest one: does this actually help the person who lands on it, or was it made to exist? Content built to serve a reader — answering the real question, anticipating the follow-up, respecting their time — performs. Content built to fill a page or hit a word count doesn't. This is a judgement about purpose, and purpose is set by the person directing the work, not by whether the typing was done by hands or by a model.

What This Means For Your Links

This matters directly for anyone buying links, because the content attached to a link is part of what makes the link work. A link sitting in a thin, generic, converged article on a site full of the same is a weaker link than the identical link sitting in genuinely useful content on a credible page — regardless of whether either was written by a human or a machine. The question to ask a content or link provider is never 'was this written by a human?' It's 'is this good — is it original, experienced, useful enough to deserve to exist and to host a link?'

That's the standard behind professionally written content created to genuinely earn its place, and it's why the authorship question is the wrong one to interrogate. The content that supports a strong link has to clear the quality bar, not the origin bar. A 'human-written' guarantee tells you nothing about whether the content is differentiated, experienced, or useful — it just tells you about the production method, which is the one thing that doesn't affect performance.

It also reframes a related debate. As AI answers increasingly mediate how people find information — the territory of how AI engines surface and cite sources — the premium on genuinely original, experience-rich content only rises, because that's the content that gets cited and remembered rather than absorbed into a generic synthesis. And it reinforces something we've said many times: even the best content still needs links to compete, which is the point of why even genuinely good content still needs promotion. Origin was never the lever. Quality, originality, and promotion are.

The Bottom Line

The 'human-written content' obsession is a comfortable proxy for a harder question. It feels safe to certify that a person typed the words, because it sounds like a guarantee of care and quality. But it guarantees neither. Humans produce thin, useless, derivative content constantly; AI-assisted work directed by someone who knows what they're doing can be genuinely excellent. The search engine evaluating the page does not know or care which it was, and neither should you. What it can tell — what it's increasingly built to tell — is whether the content is original, experienced, and useful, or whether it's just another instance of the generic average.

So stop asking whether content is human-written and start asking whether it's any good. That's the question that actually predicts whether it ranks, whether it gets indexed, and whether the link it carries does any work. For the full range of options, our full range of link building services lays out what's available, and if you want a straight answer about whether your content is genuinely earning its place or just filling space, free SEO and backlink advice is available.