Scaled content abuse is the label Google gives to one of its search spam policies: generating large numbers of pages chiefly to manipulate rankings rather than to help the people who land on them. Text written by a generative artificial intelligence (AI) model, by a team of freelance writers or by a template filled from a spreadsheet can all qualify, provided the volume is high, the added value is low and the purpose is ranking. Pages found in breach can be demoted or dropped from Google Search, and through Google's publisher rules they can also lose eligibility for Google-served ads.

The policy exists because the rule it replaced was written around automation. Once language models made fluent text almost free to produce, and light human editing could disguise machine output, asking whether a page had been automatically generated stopped being a workable test. Google moved the line from method to intent.

How the policy is applied

Google's spam policies page, last revised on August 28, 2026, describes the practice as many pages generated mainly to manipulate rankings, typically unoriginal material of little value to users. It lists five illustrative patterns: using generative AI to produce many pages without adding value; scraping feeds, search results or other sites, including through automated synonymising or translation; stitching together content from different pages; spreading output across multiple sites to hide its scale; and publishing keyword-laden pages that make little sense to a reader.

No threshold is published. Google has never disclosed a page count, publishing rate or ratio that triggers the policy, and nothing in the documentation makes volume alone a violation.

Detection runs on two tracks. The automated one is SpamBrain, Google's machine learning spam system, refreshed through announced spam updates. Its demotions arrive without notice. The other is human review: the Search Console category now labelled major spam problems cites scaled content abuse, alongside cloaking, as an example of aggressive spam. A manual action comes with a written notice and can be appealed through a reconsideration request.

Site-level judgement is built into the rater guidelines. Section 4.6.5 of the current edition, dated September 11, 2025, tells quality raters to rate sites built from scaled, valueless content as Lowest after examining several pages, even when the production method is unclear. Rater scores do not change rankings directly; Google uses them to evaluate its systems.

Remediation is blunt: the policy page tells hosts to exclude such material from Search, through a noindex directive or deletion.

The advertising link extends the consequences. Google Publisher Policies, which govern AdSense, AdMob and Ad Manager, prohibit Google-served ads on screens that violate the spam policies for Google web search. Since December 2024, Google Ads has also disapproved ads pointing to destinations removed from Search through a manual action. One finding can therefore close organic traffic, Google ad monetisation and paid acquisition through Google at once.

Google's Search Quality team writes and enforces the policy. Those exposed include publishers, affiliate operators, search engine optimisation (SEO) agencies and brands with large content programmes. Buyers have no equivalent switch: verification vendors grade inventory under their own labels, independently of Google's search verdicts.

Origin and evolution

Before 2024 the relevant rule covered automatically generated content, a long-standing violation in Google's webmaster guidelines that was qualified in April 2022 to material intended to manipulate rankings, according to Search Engine Roundtable. On April 6 that year, Google's John Mueller told an office-hours session that machine learning output was "still automatically generated content" and therefore against the guidelines, according to Search Engine Journal.

ChatGPT's public release on November 30, 2022 changed the economics. On February 8, 2023, Danny Sullivan and Chris Nelson of Google published guidance stating that appropriate use of AI was not a violation, while automation used primarily to manipulate rankings was.

In late November 2023 the marketer Jake Ward described what he called an "SEO heist": exporting the sitemap of the spreadsheet tutorial site Exceljet and using AI to publish about 1,800 articles on the same topics for the startup Causal. Ward claimed 3.6 million visits. According to figures reported by The Decoder, Causal's traffic then fell from about 610,000 visits in early October to 190,000 by mid-December 2023.

On March 5, 2024, Google introduced three new spam policies - scaled content abuse, expired domain abuse and site reputation abuse - alongside a core update and a spam update. Elizabeth Tucker, a Google director of product management, wrote that the old rule had been designed for cases where automation was clearly involved, and that the new one applied "whether automation, humans or a combination are involved." Tucker projected a 40% reduction in low-quality, unoriginal content in results; on April 26, 2024, after the rollout ended on April 19, Google reported 45%. The same update folded the helpful content system, launched in August 2022, into core ranking.

Enforcement began immediately. Ian Nuttall, who monitored 49,345 sites, counted 837 deindexed within days, about 1.7% of his sample, accounting for more than 20.7 million monthly visits and roughly $446,552 in monthly display advertising revenue, according to Search Engine Journal.

On January 23, 2025, the rater guidelines grew from 170 to 181 pages, with the low-quality sections realigned to the spam policies and a dedicated section on scaled abuse. By April 2025, raters were being told that main content produced with automated or generative AI tools and little effort could warrant the Lowest rating. On May 15, 2026, Google stated that its spam policies apply to generative AI responses in Search, AI Overviews and AI Mode included.

Why it matters for marketers

For publishers, search visits feed programmatic revenue, and the Publisher Policies tie Google's own ad demand to the same verdict.

For brands, the risk sits in their own content programmes. ClickUp's blog lost 97.6% of its traffic, from 1.19 million monthly visits in January 2025 to about 28,790 by April 2026, while its page count rose from 4,225 to 7,040, including 249 competitor comparison listicles. Neither Google nor the analyst Kamila Olexa attributed the decline to a scaled content action; it tracked a sequence of core and spam updates. Self-ranking listicles drew wider scrutiny when SISTRIX reported visibility losses of up to 49% in January 2026.

For buyers, the problem has an open-web counterpart. DoubleVerify's Fraud Lab exposed AutoBait on March 4, 2026: more than 200 made-for-advertising domains producing AI slideshows of up to 56 slides at a cost of at most $2.25 each, with up to eight banners per slide. According to the Association of National Advertisers, about 15% of programmatic spend reaches such sites. According to NewsGuard, 3,749 AI content farm sites were operating as of June 23, 2026. A Google search demotion does not, by itself, remove any of them from auctions run by other exchanges.

Limitations and disputes

Intent cannot be observed directly. The policy turns on primary purpose, yet Google judges output, and legitimate templated publishing, from property listings to sports results, shares surface features with abuse. Without a published threshold, the line between a large catalogue and a violation is drawn after the fact.

Attribution is the second problem. Algorithmic demotions carry no notice, and spam updates often overlap core updates. Marie Haynes suspected the June 2026 spam update hit sites mass-producing commodity content, then added: "I cannot prove this yet."

The evidence on outcomes conflicts. Against the SISTRIX losses, Kevin Indig of the Growth Memo found that 62% of vendor-authored listicles gained visibility after an unconfirmed January demotion, with a median traffic gain of 38%. The 45% reduction Google claimed in 2024 comes from its own evaluations, without published methodology or outside audit.

Scaled content abuse and adjacent terms

AI-generated content is a production method, not a violation. Google has permitted it since at least February 2023 when it serves readers.

Site reputation abuse, now titled the site reputation policy, concerns third-party content placed on a host domain to borrow its ranking signals. Since August 30, 2026, its enforcement works differently for users in the European Economic Area. Scaled abuse needs no host and has no regional carve-out.

Doorway abuse covers pages built to rank for narrow queries and funnel visitors elsewhere. Doorways are often mass-produced, but the policy is defined by the funnelling, not the volume. Scraping is likewise a separate policy that overlaps when copied material is republished in bulk.

Made-for-advertising sites and pink slime are advertising and media research labels, defined by monetisation practices or by impersonation of local news. Such sites may also breach Google's policy, but the classifications are made by different parties for different purposes.

Recent developments

The September 2026 spam update began on September 24 and lists scaled content abuse among the policies in scope. Google allowed up to two weeks, while its three earlier spam updates of 2026 each finished in under three days, according to the Search Status Dashboard. As of October 2, 2026, the dashboard showed no end time.

The vocabulary is spreading beyond web search. Google researchers have described a Scaled Abuse Forensics Examinerfor tracing networks of YouTube channels that distribute synthetic video, without publishing performance figures. YouTube renamed its own repetitious content rule inauthentic content on July 15, 2025.

On October 1, 2026, Google added language to its guidance on generative AI content stating that all AI-generated content should be manually fact-checked before publication, including titles, meta descriptions, structured data and image alt text. Two and a half years after the policy appeared, Google's guidance has shifted attention from what produced a page to who checked it.

Timeline

  • April 6, 2022: John Mueller says content produced with machine learning tools counts as automatically generated content under Google's webmaster guidelines
  • April 2022: Google qualifies its automatically generated content guideline to cover material intended to manipulate rankings
  • August 2022: Google launches the helpful content system
  • November 30, 2022: OpenAI releases ChatGPT to the public
  • February 8, 2023: Google publishes guidance stating that AI use is acceptable unless its primary purpose is ranking manipulation
  • November 2023: Jake Ward publicises the "SEO heist" of about 1,800 AI-written pages for Causal
  • March 5, 2024: Google introduces the scaled content abuse, expired domain abuse and site reputation abuse policies with the March 2024 core and spam updates
  • March 8, 2024: Ian Nuttall reports 837 deindexed sites from a sample of 49,345
  • April 19, 2024: The March 2024 core update finishes rolling out
  • April 26, 2024: Google reports 45% less low-quality, unoriginal content in results
  • December 2024: Google Ads begins disapproving ads to destinations removed from Search through manual action
  • January 23, 2025: The rater guidelines expand to 181 pages with a section on scaled content abuse
  • July 15, 2025: YouTube renames its repetitious content policy as inauthentic content
  • September 11, 2025: Google publishes the current edition of its rater guidelines
  • March 4, 2026: DoubleVerify publishes its AutoBait investigation
  • May 15, 2026: Google states that its spam policies apply to generative AI responses in Search
  • June 23, 2026: NewsGuard's count of AI content farm sites reaches 3,749
  • August 28, 2026: Google revises its spam policies page
  • September 24, 2026: The September 2026 spam update begins, with a window of up to two weeks
  • October 1, 2026: Google adds manual fact-checking language to its guidance on generative AI content

Summary

Who. Google's Search Quality team writes and enforces the policy through SpamBrain and human reviewers. Publishers, affiliate operators, SEO agencies and brands with large content programmes are exposed to it, and Google's ad products apply its verdicts to monetisation and paid traffic.

What. A search spam policy against generating many pages primarily to manipulate rankings, regardless of whether AI, people or a combination produced them. Consequences range from algorithmic demotion to removal through manual action.

When. Introduced on March 5, 2024, as the successor to Google's automatically generated content rule, written into the rater guidelines on January 23, 2025, extended to generative AI responses on May 15, 2026, and enforced in every spam update since.

Where. Across Google Search, including AI Overviews and AI Mode, with knock-on effects in AdSense, AdMob, Ad Manager and Google Ads.

Why. Generative AI made page production close to free, and a rule based on how content was made no longer separated useful automation from manipulation. Google shifted the test to purpose and value, a standard that remains difficult for outsiders to verify.