SpamBrain is Google's machine learning system for detecting spam in Search. It assesses pages, sites and links against the company's published spam policies and, when it finds a violation, pushes the content down the rankings, drops it from results, or cancels whatever ranking value a manipulative link was passing. Google calls it "our AI-based spam-prevention system". Its spam update documentation, last revised on December 10, 2025, says the company improves the system from time to time "to make it better at spotting spam and to help ensure it catches new types of spam."

Scale explains why it exists. Google states that "our systems find 40 billion spammy pages every day", a volume no team of reviewers could assess. According to Google, its automated systems "can detect the vast majority of spam" before it reaches top results, while "the rest of spam is tackled manually" by a review team.

How the system works

Google has never published SpamBrain's architecture, training data or thresholds. What is known comes from its documentation, two annual reports and one leak.

Placement. The system works at more than one point in the pipeline. Google's report on 2022 stated that SpamBrain could detect spam during crawling, so a page no longer needed to be indexed before it was judged spammy. Its verdicts also feed ranking. The ranking systems guide lists spam detection systems, SpamBrain among them, as the tools that deal with "content and behaviors that violate our spam policies", and describes them as continually updated.

Scope. Enforcement maps to the written policies: cloaking, doorway abuse, hacked content, scraping, scaled content abuse, expired domain abuse and link spam, among others. Links joined the system's remit on December 14, 2022, when Google said SpamBrain "can now detect both sites buying links, and sites used for the purpose of passing outgoing links."

Scores. Internal Content Warehouse API documentation that leaked in 2024, and whose authenticity Google confirmed on May 29 that year, lists three SpamBrain attributes. One, spambrainTotalDocSpamScore, is described as "the document total spam score identified by spambrain, going from 0 to 1". The other two, spambrainData and spambrainDomainSitechunkData, hold scores at host and domain level. Google spokesperson Davis Thompson cautioned against conclusions drawn from "out-of-context, outdated, or incomplete information", and the documents did not describe scoring functions.

Effects. Beyond demotion and removal, SpamBrain can neutralise links. Google's documentation is blunt: once its systems remove the effect of spammy links, the ranking benefit those links generated is lost, and it "cannot be regained."

Updates. SpamBrain runs continuously. A spam update is the announced moment when Google ships a notable improvement, logged on the Search Status Dashboard with a start and end time. These releases refresh the classifiers that sit behind SpamBrain, with rollouts running from under a day to 27 days.

No notice is sent. Unlike a manual action, a SpamBrain demotion generates nothing in Search Console and cannot be appealed. Recovery depends on the system itself: according to Google, "making changes may help a site improve if our automated systems learn over a period of months that the site complies with our spam policies."

Humans remain at the margins. Pedro Dias, a former member of Google's search quality team, described in April 2026how sites the algorithms cannot classify confidently fall into a queue for manual review. His summary of what reviewers look for: "Low effort, high monetization intent, when these two things come together, that often signals towards something wrong."

Origin and evolution

SpamBrain inherited a problem Google had been fighting since at least 2012. That April, Google launched Penguin, which the first edition of its ranking systems guide, published in November 2022, said "demoted websites with spammy linkbuilding practices". In September 2016 Penguin became part of the core algorithm and changed approach, and Google said it now "devalues spam by adjusting ranking based on spam signals, rather than affecting ranking of the whole site." Neutralising bad links rather than punishing whole domains became the approach SpamBrain later applied to link schemes.

The system itself dates to 2018, according to Google, although the company did not name it for another four years. Through 2021, Google announced two spam updates in June, another in November, and a link spam update on July 26 described as "even more effective at identifying and nullifying link spam more broadly, across multiple languages".

The name surfaced on April 21, 2022, in Google's report on 2021. "Since we first started using it, we've been updating SpamBrain constantly, and it's grown much more effective at detecting and nullifying both existing and new spam types," Google wrote. It credited the system with identifying about six times more spam sites than in 2020, alongside a 70% reduction in hacked spam and a 75% reduction in gibberish spam on hosted platforms.

The December 2022 link spam update was the first public deployment against link schemes, rolling out over two weeks in all languages. Google's next report, published in April 2023, claimed SpamBrain detected five times more spam sites in 2022 than in 2021 and roughly 200 times more than at launch. Link spam detection rose 50-fold compared with the July 2021 update, clicks to scam sites fell by half, and 99% of visits from Search were described as spam-free.

Scope widened after that. An October 2023 update extended coverage in Turkish, Vietnamese, Indonesian, Hindi and Chinese. On March 5, 2024, Google shipped a spam update alongside three new spam policies covering scaled content abuse, expired domain abuse and site reputation abuse; that rollout ended on March 20. Spam updates followed in June 2024 and December 2024, each lasting seven days. The August 2025 release ran from August 26 to September 22, a 27-day rollout.

Why SpamBrain matters for marketers

The first exposure is link spend. Since December 2022, SpamBrain has evaluated both buyers and sellers of links, putting sponsored articles and paid placements within reach. Google asks that paid and affiliate links be qualified with rel="sponsored" or nofollow. Where they are not, and the system neutralises them, the budget produces no durable ranking value, and Google describes the lost benefit as permanent.

The second is revenue. Ad-funded publishers convert search visits into impressions, so an algorithmic demotion flows straight into programmatic income. During the August 2025 update, affected publishers reported traffic declines of 70% or more. With no notice and no appeal, such losses are hard to diagnose.

The third is AI search. On May 15, 2026, Google stated that its "spam policies also apply to generative AI responses in Google Search", bringing AI Overviews and AI Mode formally under the same rules, including guidance on inauthentic mentions. Tactics built to seed brand recommendations into AI answers, such as self-ranking listicles, now sit inside the territory SpamBrain polices.

Limitations and disputes

Opacity sits at the centre of most criticism. Google does not say which policy triggered a demotion. When the March 2026 core update began three days after a spam update, practitioners struggled to separate quality signals from spam enforcement in their data.

The evidence of effectiveness is self-reported. Every SpamBrain metric in the public record comes from Google, most are relative multipliers without absolute baselines, and the "99% spam-free" figure has not been independently audited. The annual reports that carried those numbers have not been extended beyond 2022 data.

Collateral damage is disputed. After the December 2024 update, independent site owners reported sharp visibility losses while large publishers held position. Masab Farooque, whose anime and gaming site Replay Jutsu lost visibility, concluded that "anything that is not a 'brand' is 'spam' for Google". The link between those losses and the spam update was drawn by the owners, not confirmed by Google.

Misses cut the other way. In May 2025, SEO strategist Lily Ray documented self-promotional listicles surfacing in AI Overviews, arguing that "this type of spam should not be working in AI Overviews". In July 2026, fabricated earnings headlines on the Google Finance pages for Criteo and Zeta Global persisted after two global spam updates, sending readers to a WhatsApp stock club.

SpamBrain and adjacent terms

A spam update is an event; SpamBrain is the system that runs every day between those releases.

A manual action is a penalty applied by a human reviewer. It is disclosed in Search Console and can be appealed through a reconsideration request. SpamBrain demotions offer neither.

Penguin was SpamBrain's link-focused predecessor. Announced in 2012, it was, according to Google, "integrated into our core ranking systems in 2016" and now appears on the list of retired systems.

A core update changes ranking systems broadly to surface relevant, satisfying content, with no policy violation required. Core updates often run for weeks because, as Google's John Mueller put it, they "sometimes have to work step-by-step, rather than all at one time."

Recent developments

The year 2026 has brought four spam updates in seven months. The March 24 release finished in about 19.5 hours, the fastest spam update on record. A June 24 update followed, then a third that began on August 18 and applied to every language and region.

Today Google released the September 2026 spam update. According to Google, it "applies globally and to all languages" and "the rollout may take up to two weeks to complete", a longer window than any earlier release this year. The policies in scope now include back button hijacking, inauthentic mentions and undisclosed incentivised reviews, alongside cloaking, link spam and the three 2024 categories.

Eight years after Google dates its launch, SpamBrain's reach has moved from ranked links into the AI answers that increasingly sit above them.

Timeline

  • April 2012: Google launches Penguin to demote sites using spammy link-building practices
  • September 2016: Penguin becomes part of Google's core ranking systems and shifts to devaluing spam signals
  • 2018: SpamBrain enters use inside Google, according to the company, without a public announcement
  • June 23 and June 28, 2021: Google releases two spam updates
  • July 26, 2021: A link spam update targets link spam across multiple languages
  • November 2021: A spam update rolls out over eight days
  • April 21, 2022: Google names SpamBrain publicly in its report on 2021, citing about six times more spam sites identified than in 2020
  • October 19, 2022: A spam update begins and completes in about 42 hours
  • December 14, 2022: The link spam update deploys SpamBrain to detect sites buying and selling links
  • April 2023: Google's report on 2022 claims five times more spam sites detected than in 2021 and 200 times more than at launch
  • October 4, 2023: A spam update extends coverage in Turkish, Vietnamese, Indonesian, Hindi and Chinese
  • March 5, 2024: A spam update launches alongside the scaled content abuse, expired domain abuse and site reputation abuse policies; the rollout ends March 20
  • May 29, 2024: Google confirms the authenticity of leaked Content Warehouse API documentation that lists SpamBrain score attributes
  • June 20, 2024: A seven-day spam update begins
  • December 19, 2024: A seven-day spam update begins
  • August 26, 2025: A spam update begins and runs 27 days, ending September 22
  • March 24, 2026: A spam update completes in about 19.5 hours
  • May 15, 2026: Google states that its spam policies apply to generative AI responses in Search
  • June 24, 2026: A spam update begins and closes about 48 hours later
  • August 18, 2026: The third spam update of 2026 begins
  • September 24, 2026: The September 2026 spam update begins, with a rollout window of up to two weeks

Summary

Who. Google builds and operates SpamBrain inside Search. The sites it evaluates range from link sellers and scaled content farms to ad-funded publishers and affiliate operators, which makes SEO teams, link buyers and programmatic revenue teams all parties to its decisions.

What. A machine learning spam-prevention system that scores pages, hosts and domains against Google's spam policies, then demotes content, removes it from results or neutralises manipulative links, without notification or appeal.

When. In use since 2018 by Google's account, named publicly on April 21, 2022, deployed against link buying and selling on December 14, 2022, and refreshed through announced spam updates, four of them in 2026 so far.

Where. Across Google Search, from crawling through ranking, and since May 2026 formally across AI Overviews and AI Mode, which draw on the same results.

Why. Google's systems find 40 billion spammy pages a day, far more than human reviewers could handle. SpamBrain automates the bulk of that enforcement, and its decisions shape the value of link spend, the traffic behind publisher ad revenue, and the sources AI answers cite.