Guides›Google and AI

Does Google penalise AI-written content?

No. Google judges what your content is worth, not who wrote it. Here is what does count as spam, and how to use AI safely on your own business website.

By Kintsu editorial team · 4 November 2026

No. Google does not penalise a text because AI wrote it. It looks at what the text offers the reader: is it accurate, is it useful, does the author know the subject? What Google does penalise: mass-producing pages that help nobody, just to rank higher. Whether AI was involved makes no difference. An AI-assisted text that you check, enrich with what only you know and put your name to is fine. A hundred pages generated by AI in an afternoon are not.

What does Google itself say about AI content?

Google said it back in 2023: it focuses on the quality of content, not on how it is produced (Google Search Central). So AI is not banned. Spam is: using AI or other software mainly to rank higher in Google.

In a separate page on AI-written content, Google adds (Google Search Central):

AI helps you research a topic and organise your own text.
AI predicts words; it does not look up facts. A text can contain errors or invented details, so check everything before you publish.
The same goes for the title and short description that show up in Google, and for the descriptions of your photos.
Generating many pages without adding value can count as spam.

Google also has people test its search results. Their handbook (September 2025) says the same: whether AI helped write a page does not make it good or bad (Search Quality Rater Guidelines, 4.6.5 and 4.6.6). These testers do not set your ranking directly, but they show what Google wants to reward.

What does count as spam?

Since March 2024, Google has had a specific rule against scaled content abuse: mass-producing pages to rank higher, whether by AI, by people or by both (Google's spam policies). Google's examples include:

using AI or another program to create many pages that add nothing for visitors;
taking other people's texts and transforming them automatically, with synonyms or by translating them, without adding anything;
stitching together pieces of different pages;
creating pages full of keywords that make no sense to a reader.

At the same time, the question "is this content helpful?" became a fixed part of how Google ranks results (Google Search Central). This year Google has already tightened its spam checks four times: in March, June, August and September (Google Search Status Dashboard).

Three ways a small business can get it wrong:

a joiner with two hundred "windows and doors in [town]" pages, each with the same text and a different place name;
an accountant who posts an AI article every day, repeating what is already everywhere;
a hairdresser who runs a colleague's blog through a translation tool and publishes it as their own.

And your own site in two or three languages? That is not spam. Google only names automated translation of other people's content without added value. Do have your own translations checked by someone who knows the language: then you are safe, and your customers understand you.

I already have AI texts on my site. What now?

You do not need to delete them. Reread them with these questions:

Is every figure, every rule, every price correct?
Does it contain something only you know, or could any competitor have written it?
Do you have near-identical pages, such as one per town? Merge or remove them.

To see whether Google is showing your site less, use Google Search Console, a free Google service. It shows how often your site appears in Google and gets clicked. If someone at Google takes action against your site for spam, it appears in the "Manual actions" report (Google). Ask your web developer to check.

Why your experience makes the difference

Google looks for content by people who know what they are writing about, and trust matters most. Google calls this E-E-A-T: experience, expertise, authoritativeness and trust (Google Search Central). That is where you have the edge over any AI: you have the experience. Show it:

Your own work. Photos of real projects, examples with real prices, the questions customers actually ask you.
A name. Who wrote it, who checked it? Giving AI an author byline is, in Google's words, "probably not the best way" to be clear about how content was made.
Your details. Company name, address, enterprise number, phone and reviews.
Sources for figures and rules.

An example for a joiner. This is how an AI writes it:

An oak staircase is a timeless choice that brings warmth and character to any interior.

And this is how it becomes your text:

We make oak staircases to measure in our workshop in [town]. In the photo: a quarter-turn staircase we fitted last year in [town]. Expect [guide price] for a straight staircase, and [number] weeks from measuring to fitting.

If your text touches on money, health or safety, trust matters even more. Think of an accountant writing about VAT, or an installer explaining energy grants.

Google also recommends asking yourself three questions about every text: who made it, how, and why? A short note that AI helped is useful, says Google, when readers might wonder how the content was created. And the "why" should be to help people, not to attract search traffic.

So how do you use AI well?

Start from your own facts. Give the AI your prices, your process, your customers' questions and your own examples. Let it organise and phrase, not invent.
Check every fact. Every figure, every rule, every date. A wrong number costs more trust than a dull sentence.
Edit it into your own voice. Would you say it like that on the phone? Cut what is generic or already everywhere.
Put a real name on it. An author, and perhaps "reviewed by".
One good article a month beats thirty thin ones. No page per town with the same text.
Have translations checked by someone who knows the language.

What about AI Overviews or ChatGPT?

More and more people get their answers from AI: the summary at the top of Google (AI Overviews), or ChatGPT. The same rules apply there. Google says you do not need special files or code (Google Search Central). A page that answers one question clearly, with the answer at the top, stands the best chance.

The technical side is your web developer's job. Ask them:

Can Google visit my site, and are all my pages in Google?
Can ChatGPT's search crawler get in? It is called OAI-SearchBot; sites that block it do not appear in ChatGPT's search answers (OpenAI).
Are my address, services and opening hours on the site as plain text, not only in a photo or a PDF?
Is my Google Business Profile correct?
Can you check in Google Search Console that everything is in order?

Do you also have to label such texts as AI? See Do I have to label AI text on my website?

In short

Google does not penalise AI; it penalises mass production without value.
Check every fact, including the title and description.
Add what only you know: experience, examples, prices.
Put a real name on it.
Have translations checked.
Give your web developer the list above.

At Kintsu, AI writes one new text a month, translated too, and nothing goes live until you have read and approved it. If your French or Dutch is not strong enough, have the translation checked by someone who knows the language. How AI works at Kintsu.

Written with help from AI, reviewed by the Kintsu editorial team.

Sources

Accessed on 4 November 2026.

See it first, then decide

Would you rather see it than read about it?

We build your new site first, in three languages and with no setup fee. If you like it, you start a subscription.

Get your free preview