---
title: "Google changes in October 2026: what matters for SEO"
description: "Google in October 2026: a two-week spam update, mandatory fact-checks for AI content, fake authors, AI Overviews data and a multi-country filter in GSC."
date: 2026-11-08
category: SEO
tags: ["SEO", "Google", "Spam Update", "AI Overviews", "Google Search Console"]
url: https://uper.pl/en/blog/google-changes-october-2026-seo/
---

# Google changes in October 2026 Spam update, AI content, authorship, AI Overviews and Search Console

**In October 2026 Google finished a spam update that took almost two weeks instead of the usual two days. At the same time it tightened its documentation: it now calls a manual fact-check of AI content "critical" and fake authors a form of deception.** On top of that came new data on how far AI Overviews reach, the first estimate of their click-through rate and a multi-country filter in Search Console.

I started from the monthly roundup on [Search Engine Roundtable](https://www.seroundtable.com/november-2026-google-webmaster-report-42431.html). Out of several dozen items I picked the ones that change something in day-to-day work on a site, checked each in its own post, and checked the documentation and update dates against Google's own pages.

## Google's September spam update took almost two weeks

According to the [Google Search Status Dashboard](https://status.search.google.com/incidents/XhUDXP7A67iHCD2kmbVu), the update started on 24 September 2026 and finished on 8 October. The dashboard lists a duration of 13 days and 16 hours. This year's earlier spam updates all finished in under three days: March in 19.5 hours, June in just over two days, August in two days and 16 hours.

On LinkedIn Google called it a "normal" spam update for all languages and locations. While it was still rolling out, Google told Search Engine Roundtable it would take around two weeks instead of the typical two days, but gave no reason.

Barry Schwartz points to the windows in which rank trackers showed the most movement: 25-27 September, 30 September and 4-6 October. Glenn Gabe saw the 30 September change as a separate phase of the update, and some publishers on WebmasterWorld reported losing Discover and Google News traffic that day. Google has not confirmed any of these dates. If your site dropped in one of these windows, compare it with [Google's spam policies](https://developers.google.com/search/docs/essentials/spam-policies) before you change anything. I wrote about reading movement without a confirmed update in my post on [Google ranking volatility](/en/blog/google-ranking-volatility-unconfirmed-update/).

### Why there are more spam updates

At Search Central Live Deep Dive in Barcelona, Google explained that there are more spam updates because there is much more new content, and that the team now uses AI to catch more of it. Gary Illyes added that scaled content is becoming a bigger problem than link spam, and repeated that Google filters around 40 billion spam pages a day. All of this is secondhand, from John Campbell's (We Are Roast) recap, which Search Engine Roundtable summarised.

For an example of the scale involved, see my post on [Facebook Answers pages in Google's index](/en/blog/facebook-answers-google-ai-pages/).

### More manual actions

On 7 October Marie Haynes wrote that she went from essentially no requests for help with manual actions to several a week. They cover thin content, pure spam and scaled content, and most of the affected sites are, in her words, legitimate long-standing businesses. In November she added that reconsideration requests usually take weeks, sometimes months, and often several attempts. Daniel Foley Carter reports the same rise.

Both consultants describe what they see among their own clients, and Google has not published any numbers.

## AI content, main content and authorship in Google's documentation

In early October Google changed three parts of its documentation, and all three are about the same thing: who is responsible for the content and whether they checked it.

### Manual review of AI content

In its [guidance on using generative AI content](https://developers.google.com/search/docs/fundamentals/using-gen-ai-content) Google now says it is **critical** to manually fact-check and review all AI-generated content before publishing. The reason it gives is that generative models don't retrieve facts, they predict a likely sequence of words based on their training data.

The requirement also covers metadata: titles, meta descriptions, structured data and image alt text. The same page warns that generating pages at scale without value for readers may violate the policy on **scaled content abuse**. Google says the change brings the documentation in line with what it has been saying at conferences for a long time.

### Quality of the main content

The page on [helpful, reliable, people-first content](https://developers.google.com/search/docs/fundamentals/creating-helpful-content) gained a section on **main content**. Google writes that its quality is one of the most critical factors in assessing a page. Main content is any part of the page that directly helps it achieve its purpose:

- primary text and media,
- interactive features,
- user-generated contributions,
- tabbed or expandable sections,
- page titles and headings.

A second new section lists four attributes of main content that Google trains its Search Quality Raters to evaluate: **effort, originality, talent or skill, and accuracy**. For YMYL topics (health, finance, safety) the content must be highly accurate and consistent with established expert consensus. Barry Schwartz shortens this to "EOT/SA", but Google does not use the acronym.

### Fake authors are deception

In the same document, in the part about who created the content, Google says plainly that fabricating creator profiles is a form of deception. Its examples are AI-generated headshots, made-up names and false credentials. Any form of deception makes a page untrustworthy to both users and Google's automated quality systems, and is itself a signal of a low-quality page.

At the same time Google encourages accurate authorship information, such as a byline.

### What this means in practice

None of these rules is new, but the tone has changed: the documentation now says "critical" and "deception" where it used to leave room for interpretation. In audits I rarely see a problem with the language model itself. Far more often the problem is a missing editor, when texts go live without anyone checking the facts and without an author who answers for them.

I would not assume these changes describe exactly what the September spam update targeted. Barry Schwartz connects the two, but Google has not confirmed it. For more on using AI in content, see my posts on [AI content generation and SEO](/en/blog/ai-content-generation-and-seo/) and on [how Google treats AI and human content](/en/blog/ai-vs-human-content/).

## How long Google takes to react: Gary Illyes' numbers

In Barcelona Gary Illyes also showed typical and slowest times for individual Search processes. He said it was an exercise to see whether the audience could relate to numbers pulled from internal analysis. They are not a guarantee, but the table helps when a client asks when they will see the effect of a fix.

| Process | Typical | Slowest |
|---------|---------|---------|
| Discovering a new URL | about 20 hours | weeks or never |
| Reading a sitemap | about 24 hours | up to 14 days or never |
| Indexing (end to end) | about 1.5 hours | months or never |
| Canonical change | 1-3 weeks | not given |
| Site move | 1-3 months | 6 months to over a year |
| Manual action removal | 1-2 weeks | 4-6 weeks |
| Effect of fixes after a spam update | 1-2 weeks | months |
| Recovery after a core update | 3-6 months | 6-12 months (the next core update) |

These times add up, because each stage waits for the previous one. If you are fixing a site after a core update drop, expect the improvement around the next big update. I cover this in more depth in my post on an [update-proof SEO strategy](/en/blog/update-proof-seo-strategy/).

## AI Overviews: reach, links and click-through rate

### Reach depends on who is counting

The Search Engine Roundtable headline says over 80% of searches have AI Overviews, but that is Barry Schwartz's summary of several tools at once. The studies themselves give these numbers:

| Source | Result | Scope |
|--------|--------|-------|
| SE Ranking | 82.5% | US branded searches |
| seoClarity | over 62% | all queries, 55% to 84% depending on industry |
| Semrush | no number in the post | a chart in its Sensor tool |

The data supports a more careful conclusion than the headline: AI Overviews appear on most queries, and on the vast majority in some industries.

### More links in AI Mode and AI Overviews

On 15 October Nick Fox of Google announced another change to links in AI answers:

- more links embedded directly in the response text,
- a new section at the top with one or two links when the query is about completing a task,
- a redesigned "Dive Deeper" section with thumbnails,
- news carousels rolled out globally, highlighting Preferred Sources.

Google says the changes should make it easier to click through to source pages, but it shared no click data and only said it will add more metrics over time.

### How many clicks AI visibility brings

The only public estimate comes from a small experiment by Cyrus Shepard, who opted one site out of AI features with the "Search generative AI" control in Search Console settings. AI impressions made up about 8% of that site's total impressions, and he estimated their CTR at about **0.9%**.

After the opt-out, regular organic clicks dropped too, which Shepard stresses. He also notes that the sample is small, and that AI Overviews still appeared in results, just without his site. Barry Schwartz adds that nobody else has published similar data. I treat it as a hypothesis to test, not as an argument for opting out. The control does not remove a site from regular results and does not block model training; that is what Google-Extended is for. For the wider picture, see [Do AI answers kill traffic?](/en/blog/do-ai-answers-kill-traffic/).

## Search Console: multi-country filter, AI report still without clicks

In the Search Console performance report you can now select several countries at once and see their combined traffic. For businesses that work in several markets this saves time, since until now the data had to be added up by hand or in a spreadsheet.

The [generative AI report in Search Console](/en/blog/search-console-ai-search-reports/) still shows only impressions from AI Overviews and AI Mode, broken down by page, country, date and device, with a filter for text and multimodal searches (Lens, Circle to Search). There are no clicks and no queries. In early November Gary Illyes wrote on LinkedIn that the team wants to share more data but has nothing to announce yet.

I cover the basics of the tool in my [Google Search Console guide](/en/blog/google-search-console/).

## What else was in the roundup

I wrote separate posts on a few other topics from this period:

- [Google's guidance on hyphens in URLs](/en/blog/hyphens-in-urls-google-guidelines/),
- [the robots.txt rules Google actually supports](/en/blog/google-robots-txt-what-you-can-put-in-it/),
- [taking a site offline with a 503 while patching a PHP vulnerability](/en/blog/google-503-take-site-offline-php-patch/),
- [Universal Cart, a shopping cart in Search](/en/blog/google-universal-cart-shopping-cart-in-search/),
- [Google Posts higher in the local panel](/en/blog/google-posts-higher-local-panel/).

The roundup also covers new Google Business Profile reports and features, a Gemini model in AI Mode, a citation cards test in AI Overviews and Google's Q3 results. I don't go into those, because they don't change day-to-day work on a site.

## Summary

Google now talks about quality directly: fact-checking AI content is "critical", a made-up author is "deception", and the quality of the main content is one of the most important factors in assessing a page. Spam updates are more frequent and longer, and according to Google AI helps with them.

AI Overviews appear on most results, but we still know little about clicks. The only estimate, a 0.9% CTR, comes from a single site. Fixes also take time: according to Illyes, the effect shows 1-2 weeks after a spam update and up to six months after a core update.

What to do now: check that content written with AI help goes through a real editor, that your authors are real and described, and compare any drops from late September and early October with the update calendar before you change anything on the site.

<FaqBlog
  questions={[
    {
      question: 'When did the Google September 2026 spam update finish?',
      answer: 'It started on 24 September 2026 and finished on 8 October 2026. It took 13 days and 16 hours, while earlier spam updates in 2026 finished in under three days.'
    },
    {
      question: 'Does Google forbid using AI to create content?',
      answer: 'No. Google requires AI-generated content, including titles, meta descriptions, structured data and alt text, to be <strong>manually fact-checked</strong> before publishing. The problem is the lack of review and mass pages without value, not the tool itself.'
    },
    {
      question: 'Can I sign articles with a made-up author?',
      answer: 'Google treats fabricated creator profiles, including AI-generated headshots, made-up names and false credentials, as deception. According to the documentation this is a signal of a low-quality page and makes it untrustworthy to both users and automated systems.'
    },
    {
      question: 'How long does it take to recover traffic after a Google update?',
      answer: 'According to the numbers Gary Illyes shared, the effect of fixes usually shows 1-2 weeks after a spam update and 3-6 months after a core update, sometimes only at the next big update. Removing a manual action usually takes 1-2 weeks.'
    },
    {
      question: 'Will I see AI Overviews clicks in Search Console?',
      answer: 'No. The generative AI report shows only impressions, broken down by page, country, date and device. Google says it wants to share more data but has not given a date. I cover the details in my <a href="/en/blog/search-console-ai-search-reports/">post on AI reports in Search Console</a>.'
    }
  ]}
  heading="Frequently Asked Questions"
  id="faq"
/>

## Sources