In short
On 1 October 2026 Google revised its documentation on content created with AI. The position itself has not changed - AI is a permitted tool and Google judges the output, not the method - but the wording around scaled content abuse got sharper. The line is added value: content produced in volume that does not add anything beyond what already exists is the thing the policy targets, regardless of whether a human or a model wrote it.
What actually changed
Google did not announce a new rule. It tightened the language of an existing one, which in practice is more useful than a new rule because it tells you how the existing one is being read.
The clarification is that volume plus absence of added value is the problem, and that the question of who or what produced the text is not the test. A thousand near-identical pages written by freelancers fall under the same policy as a thousand generated ones.
Why this matters more for e-commerce than for blogs
Most coverage of this topic treats it as a question about articles. In practice the bigger exposure is in product catalogues.
A store that generates descriptions for several thousand products, all following the same template with the attributes swapped, is producing scaled content whether or not anyone thinks of it that way. The same applies to location pages that differ only by place name, and to category pages built by permutation.
That is not an argument against using AI on a catalogue - we do it, at volume. It is an argument for the output being specific: real attributes, real differences, information the manufacturer's sheet does not already state.
What to check
- Sample your own product descriptions. Take ten at random. If you can swap the product name between them and nothing reads wrongly, they are templated, not written.
- Look for near-duplicate pages. Location and category pages that differ by one word are the most common exposure.
- Check your product feed as well as your pages. Feed content is also content.
- Ask what a page adds. If the honest answer is "it exists so we rank for the term", that is the page the policy describes.
The reassuring part of this update is what it does not say. There is no requirement to label AI-assisted content, and no penalty attached to the method. The judgement is on whether the page is worth the reader's time.
Sources
- Google Search Central: AI-generated content guidance
- Google Search Central: Spam policies - scaled content abuse
Want to know what an AI agent sees on your site?
We check exactly that as part of an audit: crawler access, what is in the initial HTML, and the layer of information an agent can actually cite. Free SEO & AEO report → or talk to us →