Does Google penalise content written with AI?

No, not for being written with AI. Google says plainly that using AI appropriately isn't against its guidelines. What it does penalise is content made at scale to manipulate search rankings rather than to help people, however it was produced. That cuts both ways: a careless run of AI-generated pages can be caught by the policy, and so can a careless run of human-written ones. Most advice on the subject gets one half or the other wrong.

This is part of our series on marketing myths, checked, in which we test what small and medium businesses are told against what the people who run the systems actually publish.

What Google has said about AI content

Google's position fits in the headline of a Search Central post from February 2023: "Rewarding high-quality content, however it is produced". Danny Sullivan and Chris Nelson wrote it on behalf of the Search Quality team. The post's FAQ is short and direct: "Appropriate use of AI or automation is not against our guidelines." It also says where the line falls: using automation, including AI, to generate content "with the primary purpose of manipulating ranking in search results is a violation of our spam policies."

That post is no relic. Google's current guide to its generative AI features, updated in July 2026, still links to it, and says: "If you're using generative AI tools to assist in content creation, be sure that your work meets the standards of the Search Essentials and our spam policies."

The policy that actually applies

The rule that catches bad AI content is Google's policy on scaled content abuse, part of its spam policies, which were last updated on 28 August 2026. It defines scaled content abuse as "when many pages are generated for the primary purpose of manipulating search rankings and not helping users", typically "creating large amounts of unoriginal content that provides little to no value to users, no matter how it's created."

Its examples include:

• "Using generative AI tools or other similar tools to generate many pages without adding value for users"
• scraping feeds or search results to generate many pages, including by "synonymizing, translating, or other obfuscation techniques"
• "Stitching or combining content from different web pages without adding value"

The same page sets out what happens next: Google detects policy-violating practices "through automated systems and, as needed, human review that can result in a manual action", and "Sites that violate our policies may rank lower in results or not appear in results at all."

Three things decide it, and none of them is who typed it

Read the policy closely and it turns on three things.

Scale. It's about "many pages", not a single article. One AI-assisted guide isn't what the policy describes.

Purpose. The test is whether pages were made "for the primary purpose of manipulating search rankings". Google's guidance on helpful, people-first content frames the same point as a question of "Why": content made "primarily to help people" is what its systems seek to reward, and among its warning signs it asks: "Are you using extensive automation to produce content on many topics?"

Value. The examples all end the same way: "without adding value for users", "little to no value". That applies to a person copying and rewording other sites just as much as to a machine.

Two made-up cases show the difference. In the first, a business uses an AI tool to draft one guide, and then a person who knows the subject rewrites it, checks every fact against its source and adds what only they know. Nothing in the policy is aimed at that. In the second, a business pays someone to write three hundred near-identical pages, one per town, each saying the same thing with the town's name swapped in. That fits the policy's description of scaled content, and no AI was anywhere near it.

Where the usual advice goes wrong

"Google penalises AI content." It doesn't, as such. Google's own FAQ says appropriate use of AI isn't against its guidelines, and the scaled content policy says it applies "no matter how it's created", which works in both directions.

"AI content is fine, publish as much as you like." That's wrong too. The first example Google gives of scaled content abuse is using generative AI tools to generate many pages without adding value. Volume without value is exactly what the policy is written against.

"You must label anything written with AI." Google doesn't say that. Its AI content FAQ says AI or automation disclosures are useful where someone might think "How was this created?", that you should consider adding them "when it would be reasonably expected", and that giving AI an author byline "is probably not the best way" to make its role clear. Its advice on bylines is to use accurate ones where readers would reasonably expect them.

"Google can detect AI writing and marks it down." Google doesn't say whether it tries to detect AI authorship, only that it looks for policy-violating practices. Anyone who tells you exactly what Google's systems detect is claiming more than Google has published.

What Bing says

Microsoft draws a similar line. Bing's webmaster guidelines list "Automatically Generated Content at Scale" among the things to avoid: "Large-scale content generated without oversight, quality control, or editorial review often lacks usefulness, accuracy, and originality, and may be excluded from indexing." The words that matter are oversight and editorial review, not the tool.

What to do

1. Use AI to assist, not to replace judgement. A person who knows the subject should check every fact against its source, and add what only your business knows: your prices, your process, your examples.
2. Brief by what the page must answer. A piece is finished when it answers the question, not when it reaches a length, and our guide to content strategy on a small team shows how to brief for that.
3. Publish fewer, better pages. If a page wouldn't be worth publishing without search engines, don't publish it for them.
4. Use accurate bylines. Say how a piece was made where a reader would reasonably expect to be told and wouldn't otherwise assume it: an image that looks like a photograph but isn't, say, or a review that reads as first-hand when it isn't. Google's own wording is to consider a disclosure "when it would be reasonably expected"; it doesn't ask for a label on every page a tool helped with.
5. Remember the rest of search still applies. Whoever wrote a page, the same foundations decide whether it's found, and we cover them in our pieces on whether SEO is dead and why a website isn't showing up on Google.

A quick check

• A person who knows the subject checks every fact before a page goes live.
• Each page adds something only we could write.
• We are not producing pages mainly to catch searches.
• We have not published many near-identical pages.
• Our bylines are accurate.
• Where a reader would not otherwise assume how something was made, and would expect to be told, we say.

What nobody has published

• Detection. Google doesn't publish whether or how it identifies AI-written text, and we found no primary source that settles it either way.
• Outcomes. We found no published data from Google or Bing comparing how AI-assisted and human-written pages perform in search.
• Age of the guidance. Google's main statement on AI content dates from February 2023. Its July 2026 guide still links to it as current, and its spam policies were updated in August 2026, so we treat both as Google's present position.

So, does Google penalise AI content?

False, and so is the opposite. Google doesn't penalise content for being written with AI. It penalises pages made at scale, without value, to manipulate rankings, whether a machine or a person produced them. The honest question to ask of any page isn't "was AI involved?" but "who is this for, and what does it add?"

If you'd like a plan for what to publish, with briefs for whoever ends up writing it, that's our content plan and first briefs, at a fixed £2,500.