Contact

WritingGuide

Google's guide to generative AI content: the October 2026 update

By · Published · Updated

Sources checked on

How this was made. AI agents drafted this piece, a separate checker agent checked its sources and claims, and Rues read and approved it. How we work with AI

Most of what we publish starts with an AI agent. A separate checker agent reviews it, and a person decides what goes live. So when Google changed its guide on using generative AI content, we read the new version line by line and held our own site up against it. Below: what the guide says today, what we could confirm about the change, what we found on our own pages and how we fixed it.

What changed on 1 October 2026?

Google's list of documentation updates has this entry under October 2026:

"What: Updated the using generative AI content guide with information from the Search Quality Raters guidelines."

"Why: To get our documentation in sync with our presentations we use at our developer events."

The guide itself shows "Last updated 2026-10-01 UTC". Google did not publish a line-by-line diff, and we could not open an archived copy of the earlier page, so we will not tell you which sentences are new. The part that matches the changelog is a paragraph that points to two sections of the rater guidelines:

"You might find value in looking at the Search Quality Raters guidelines on how to evaluate both scaled content abuse (section 4.6.5) and main content created with little to no effort, little to no originality, and little to no added value (section 4.6.6)."

The same paragraph adds a limit worth keeping in mind: these guidelines "are not a guide to ranking first in Google", and the raters' ratings "don't directly influence ranking".

What do sections 4.6.5 and 4.6.6 say?

The rater guidelines PDF linked from the guide is dated 11 September 2025. Section 4.6.5 opens with one sentence that sums up the risk:

"Creating an abundance of content with little effort or originality with no editing or manual curation is often the defining attribute of spammy websites."

Its first example is "using automated tools (generative AI or otherwise) as a low-effort way to produce many pages that add little-to-no value for website visitors". Raters are told to give such sites the lowest rating "no matter how they are created".

Section 4.6.6 gives the lowest rating when all or almost all of the main content is "copied, paraphrased, embedded, auto or AI generated, or reposted from other sources with little to no effort, little to no originality, and little to no added value". All three conditions sit in the same sentence. The section also closes a common loophole: such pages get the lowest rating "even if the page assigns credit for the content to another source."

Google's page on helpful content describes effort in the same terms. It names "using generative AI to produce large amounts of text without manual oversight or curation" as little to no effort, and says "attribution or giving credit to other sources doesn't replace the need for original effort."

What does the guide ask site owners to do?

The current guide has three parts.

Stay clear of scaled content abuse. The opening paragraph allows AI for research and structure, and then draws the line: "using generative AI tools or other similar tools to generate many pages without adding value for users may violate Google's spam policy on scaled content abuse." The spam policy itself adds that the problem holds "no matter how it's created."

Check every fact by hand. The guide explains why: "generative models don't retrieve facts, but predict a likely sequence of words based on their training data." Its instruction: "It is critical to manually factcheck and review all AI-generated content for accuracy and trustworthiness before publishing." The review covers more than the body text. It names title elements, meta descriptions, structured data and image alt text, because all of these can appear in search results.

Tell readers how the content was made. The guide suggests "adding information on how your content was created in a way that makes sense for your audience, such as by providing more background information on how automation was used and adding image metadata." For online stores the rules are stricter: under Merchant Center policy, AI-generated images "must contain metadata using the IPTC DigitalSourceType TrainedAlgorithmicMedia metadata", and AI-generated product titles and descriptions must be specified separately and labeled as AI-generated.

The helpful content page turns this into questions. Under "How", it asks whether the use of automation is "self-evident to visitors through disclosures or in other ways", whether you explain how it was used, and why it was useful. Under "Who", it asks whether pages carry a byline and whether that byline leads to more information about the author. It also warns that "fabricating creator profiles (such as by using AI-generated headshots, made-up names, or false credentials...)" is deception.

How does our setup line up with it?

Our setup has four roles. A lead session splits the work and writes a brief. An agent writes the draft on its own git branch. A checker agent with no edit tool reviews it once and answers APPROVE or FIX. A person decides what goes live. We describe the full setup in Running AI agents with a checker.

Here is how that maps to the guide:

Google asks forWhat we doStatus
Manual fact check before publishingTwo steps: the checker agent matches every claim to its source and asks whether any number, date or source is made up; then Rues reads the post before it goes liveIn place
Review of metadataThe page description and the short answer go through the same review, since they become the meta description and part of the structured dataIn place
No pages without added valueEach post answers a question from our own work; the writing section has five posts as of 8 October 2026In place
SourcesEvery post ends with a source listIn place
Context on how content was madeEvery post has a "How this was made" note under the byline that links to How we work with AI; added on 8 October 2026In place
A byline that leads to the authorThe name in the byline links to About Rues; added on 8 October 2026In place
AI image metadataOur share images are rendered from an HTML template in a browser, with no generative model involved, so the IPTC label does not apply today; the rule for AI images is on our method pageNot needed yet

Our audit found two gaps, in the "How" and "Who" questions. We closed both on 8 October 2026: each post now says how it was made, and the byline leads to a page about the author. This post was made the same way: an agent drafted it, every quote was compared with Google's live pages on 8 October 2026, and the decision to publish was Rues's.

What should you check on your own site?

  • Is there any page that exists only because a tool could produce it?
  • Did a person check every number, date and quote against its source before publishing, including the title, meta description, structured data and alt text?
  • If a reader asks "How was this created?", does the page answer?
  • Does each byline lead to a real person with real background? No invented names, no AI headshots.
  • Do AI-generated images carry IPTC DigitalSourceType metadata? For Merchant Center product images it is required.
  • Do you keep an llms.txt? Google's guide to generative AI features says Google Search ignores such files, and that keeping one "will neither harm nor help" your visibility in Google Search. We keep ours for other systems and say so in the file.

If you run content through AI agents and want to compare notes, write to us.

Sources

Want to talk about this for your brand?

Write about your project and what you want to do. The message goes straight to Rues.

Open the contact form

Emailinfo@ruesandora.com Telegramruesandora X@ruesandora0