SpamBrain is Google's AI-based system for detecting spam in search results. Google has said it launched SpamBrain in 2018 and first named it publicly in 2022. It identifies spammy sites and, since a December 2022 link spam update, has also been used to detect both sites buying links and sites used to pass outgoing links. Improvements to SpamBrain are often what Google means when it announces a spam update.
What it does
SpamBrain looks for patterns associated with spam: auto-generated pages, hacked content, cloaking, link schemes and other policy violations. Because it is a machine learning system, it can generalize from known spam to new variations rather than relying only on fixed rules.
SpamBrain and links
For links, Google describes SpamBrain as neutralizing the effect of unnatural links. That means links identified as spam simply stop counting. Sites that relied on them lose that benefit, which is why link spam updates often show as ranking declines without any manual action.
What to do about it
There is no way to check whether SpamBrain has flagged your site. The useful response is to keep well inside Google's spam policies and to treat any tactic that depends on hiding something from Google as a liability.
- Avoid buying links or selling links without rel attributes
- Moderate user-generated content
- Secure the site against hacking and injected spam
- Review programmatic pages for genuine value
Example
Hacked pages on a forgotten subdomain
Hypothetical events app Gatherwick discovers an old marketing subdomain running an outdated CMS that was hacked to host spam pages. After the subdomain's pages vanish from results, the team removes the subdomain, fixes the server, and checks Search Console's Security issues report.
Common mistakes
Assuming AI systems miss clever tactics
Machine learning detection is designed to catch new variations. Tactics that work for a while can stop working suddenly.
Questions
Not exactly. SpamBrain is the system; spam updates are announced improvements, often to SpamBrain.
