Google's Spam Policies on Scaled Content Abuse: What the Policy Targets

Google publishes its spam policies for Google web search on Search Central. Scaled content abuse is one of them. Google defines it as being "when many pages are generated for the primary purpose of manipulating search rankings and not helping users."

The first example Google lists is direct: "Using generative AI tools or other similar tools to generate many pages without adding value for users."

That example gets read as a ban on AI writing. It isn't one. Google's separate guidance on AI-generated content says "Appropriate use of AI or automation is not against our guidelines." Our digest of Google's guidance on AI-generated content covers that document in full.

What it means: the policy is about intent and value at scale. The production method is secondary. For a short definition you can share with a team, see scaled content abuse.

Read the definition closely and three conditions stand out:

• Many pages. The policy is about scale, not a single thin article.
• Primary purpose of manipulating rankings. The pages exist to capture search traffic.
• Not helping users. The pages add no value a reader would miss.

The policy applies "no matter how it's created." A content farm of human writers, a template that spins city names, and a one-click AI generator can all fit the pattern. An AI-assisted page that answers a real question well does not.

Google's spam policies don't say which system detects scaled content, or whether action is manual or algorithmic in a given case. That matters in practice. A manual action shows up in Search Console. An algorithmic removal may not show up anywhere. Our comparison of manual action vs algorithmic deindexing explains how to tell them apart.

The next move: judge a publishing plan by asking whether each page would earn its place if search traffic didn't exist.

OtterlyAI, a vendor that sells AI visibility tracking, tested the volume shortcut in 2026 and published the results. This is one experiment, not a Google statement. Treat it as an illustration of how the policy can play out.

The method

Otterly launched two brand-new domains about capybaras, capybaraverse.com and capybara-world.com. Each got roughly 1,000 AI-generated posts in one click. capybara-world.com showed an author box presenting a fictional wildlife biologist with more than 15 years of field experience. capybaraverse.com had no author.

Otterly tracked Google Search Console, Bing Webmaster Tools, AI citations across seven engines, and AI crawler visits for about four months. The write-up was last updated on 2026-08-05.

What happened in Google

Both sites were removed from Google's index. Otterly reports zero manual actions in Search Console on either domain.

• capybara-world.com lasted about 14 days. It peaked at 2,426 daily impressions on April 5, then fell from 1,629 to 15 on April 9.
• capybaraverse.com lasted about 90 days. It built to more than 21,000 weekly impressions by mid-May, then fell from 859 to 85 daily impressions on June 26.

The site with the fictional expert went first. That fits Google's position that E-E-A-T isn't a specific ranking factor, covered in our digest of Google's helpful content guidance and E-E-A-T. The full chart breakdown sits in our guide to mass-publishing AI content.

Losing Google did not end the sites' AI visibility. It moved it.

• Bing never deindexed either site. Combined Bing impressions grew every month, to 119,935 in July.
• Otterly tracked 100 citations for capybara-world.com in April and 1,666 in July 1 to 27.
• Of 3,934 tracked citations for that site, 2,718 (69%) came from Microsoft Copilot, which grounds on Bing.
• Google AI Overviews, AI Mode and Gemini almost never cited the sites. ChatGPT citations were effectively zero, even though ChatGPT's crawlers kept visiting.

The Google side follows from Google's own rule. Its AI features documentation says a page must be "indexed and eligible to be shown in Google Search with a snippet" to appear as a supporting link in AI Overviews or AI Mode. Once the index dropped the pages, Google's AI surfaces had nothing to cite. See our digest of Google's AI features documentation.

What it means: a spam action in Google Search removes you from Google's AI answers too. Other engines draw on other indexes, so each one needs checking on its own.

Otterly published its data, which is more than most GEO advice does. It is still one pair of sites.

• One site per arm. The author box was not the only difference. Domain names and everything tied to a separate domain differed too.
• No human-written control. The test can't say whether Google reacted to AI text or to a thousand pages appearing overnight on a domain with no history. The policy targets the second.
• A low-stakes topic. Capybara questions have little competition. Copilot citing thin pages here says little about commercial categories.
• Prompt-dependent counts. Microsoft's own AI Performance report credited capybara-world.com with 42,817 citations, against 3,934 in Otterly's tracker. The time windows may not match.
• Vendor-run. Otterly sells tracking, and so does 99Visibility. That is a reason to want replication, not a reason to dismiss the data.

It does not prove that author boxes hurt. It also does not show that Copilot citations offset losing Google.

The policy is clear enough to act on. Build the review in before publication, not after impressions collapse.

1. Test each template against the definition. Ask whether the pages exist mainly to capture rankings. If the honest answer is yes, cut or rebuild them. Our pre-publish review for scaled AI content turns this into a checklist.
2. Add value a reader can see. Specific facts, prices, data or answers that the page states plainly. Volume alone is not a strategy, as our answer on whether publishing more content helps AI visibility explains.
3. Watch index status engine by engine. Check Google Search Console and Bing Webmaster Tools separately. A healthy Bing index tells you nothing about Google.
4. Don't wait for a warning. Otterly's sites got none. Track daily impressions on new sections and investigate sharp drops fast.
5. Confirm crawlers can reach you. Access comes before indexing. The AI Crawler Checker requests a page as each AI crawler and shows which ones your firewall turns away.

Nobody controls whether an engine cites you. You do control whether your pages fit the pattern Google describes as spam.

Does Google penalize AI-generated content? Not for being AI-generated. Google's guidance says appropriate use of AI or automation is not against its guidelines. It acts against scaled content abuse: many pages produced to manipulate rankings without adding value. In Otterly's 2026 test, Google removed both 1,000-post AI sites from its index, and Otterly reports no manual action on either. The volume and lack of value were the pattern, not the model.

Will I get a warning before Google removes scaled content? Not necessarily. A manual action appears in Search Console, but algorithmic removal may not. In Otterly's experiment both sites disappeared from Google with zero manual actions reported. Google does not say which system removed them. The practical response is to watch impressions on newly published sections closely and treat a sudden drop as a signal to investigate.

Can AI engines still cite a site Google has deindexed? Yes, if they draw on a different index. In Otterly's test, Microsoft Copilot, which grounds on Bing, kept citing the sites because Bing kept them indexed. Google AI Overviews, AI Mode and Gemini almost never did, since Google requires a page to be indexed and snippet-eligible to appear as a supporting link. ChatGPT citations were effectively zero.

Does an expert author box protect scaled AI content? Nothing in the evidence suggests it does. In Otterly's test the site with a fictional wildlife biologist author box was deindexed after about 14 days, the site with no author after about 90. One site per arm can't prove author boxes hurt. It does match Google's statement that E-E-A-T isn't a specific ranking factor, so there is no score an author box raises.

Is programmatic or templated content always scaled content abuse? No. The policy targets pages generated for the primary purpose of manipulating rankings and not helping users. Templated pages that each carry distinct, useful information for a real reader fall outside that definition. The risk rises when pages differ only cosmetically, exist to capture search variants, or would have no reason to exist without search traffic.