Google March 2026 Spam Update & AI Patent: What Actually Changed

Google released its March 2026 spam update on March 24 at 19:00 UTC and marked it complete on March 25 at 14:30 UTC: 19 hours and 30 minutes, according to the Google Search Status Dashboard↗. The March 2026 core update started on March 27 and completed on April 8. Two different updates, three days apart, which you read differently in Search Console.

The first version of this article, published on March 30, had three substantive errors: it gave other dates for the update, attributed to Google a patent application that belongs to another company, and described detection mechanisms that no Google source describes. This version corrects them, with a source for every claim. We kept the address, because a correction belongs where the mistake was.

What happened, with the official dates

The dashboard announcement says only this: "Released the March 2026 spam update, which applies globally and to all languages. The rollout may take a few days to complete." The next day, the same page noted: "The rollout was complete as of March 25, 2026."

Neither the dashboard nor Google's documentation on spam updates names targets for this update. Any text that says exactly what March "hit", including the first version of this article, describes an assumption.

The same dashboard lists four spam updates in 2026 up to October 29. March was the shortest:

Times as recorded in the dashboard's incident history, read on October 29, 2026.

What a spam update targets, in Google's words

Google's documentation on spam updates says its automated spam detection runs all the time, and an update is a notable improvement to it: "While Google's automated systems to detect search spam are constantly operating, we occasionally make notable improvements to how they work." The system it names is SpamBrain: "For example, SpamBrain is our AI-based spam-prevention system."

The reference for what counts as spam is Google's spam policies↗, updated on August 28, 2026. For content produced at volume, the policy is scaled content abuse: "many pages are generated for the primary purpose of manipulating search rankings and not helping users", content of little value to the reader "no matter how it's created". The example that concerns AI is about volume without value: "Using generative AI tools or other similar tools to generate many pages without adding value for users".

The policy the first version presented as the main target is the site reputation policy, known in the trade as "parasite SEO". It concerns third-party content published on a host site "mainly because of that host's already-established ranking signals", and hosting third-party content is not a violation in itself: "Having third-party content alone isn't inconsistent with the site reputation policy". Under the current policy, a part of a site suspected of breaking it goes to a human review ("a site will be subject to human review"), and the outcome depends on where the searcher is. Outside the European Economic Area, the pages may receive a manual action. For users within the EEA, "the relevant pages may be categorized as separate from the main domain but won't be subject to the impact of manual action": those pages may rank on their own merits, apart from the rest of the domain.

On AI, Google wrote in February 2023 that "not all use of automation, including AI generation, is spam" and that "Appropriate use of AI or automation is not against our guidelines." Its current page on generative AI content, updated on October 1, 2026, draws the line at volume without value: "using generative AI tools or other similar tools to generate many pages without adding value for users may violate Google's spam policy on scaled content abuse". What a new text should add to what already exists is the subject of our article on AI-generated content and information gain.

Spam update or core update:How to read a March drop

If you lost traffic at the end of March, two windows sit close together on the same chart. Read them in this order:

• The manual actions report first. Google's spam policies say violations are found "both through automated systems and, as needed, human review that can result in a manual action", and a manual action shows in the Manual actions report in Search Console. An empty report tells you that you don't have a manual action; it tells you nothing about the automated systems.
• A clean reference week. Google's core updates guide recommends "waiting at least a full week after a core update completes before analyzing your site in Search Console", then comparing with "a week before the core update started rolling out". For March 2026, the clean week before is March 17–23, because the spam update was already rolling out on the evening of March 24; a comparison with March 20–26 mixes the two windows. The analysis starts after April 15, with the week of April 9–15.
• The daily curve, not the monthly total. A sharp change on March 24–25, or right after, falls in the spam update window; one that starts after March 27 and runs until April 8 overlaps the core update. The data shows you where to look, not the cause: seasonality, your own changes and other search changes sit on the same chart.
• The size of the drop. The same guide separates a small drop, from position 2 to 4, where it advises against changes ("we recommend avoiding making changes to content that's already performing well"), from a large one, from 4 to 29, which calls for a deeper assessment. Read each search type separately: web, images, video, news.

The Search Console steps are in our advanced Search Console guide, and the 2026 core updates in our analysis of the core update and information gain.

Patent US12536233B1:What it says and what it doesn't

The first version of this article described a Google patent for detecting AI-written text. The number it gave belongs to another company's application, about something else (see the corrections at the end). A Google patent from 2026 that does deal with AI and search results is US12536233B1. We read it in full on October 29, 2026, on Google Patents↗, so you can judge it by its text.

The facts: its title is "AI-generated content page tailored to a specific user", the assignee is Google LLC, the priority date is July 25, 2024, the application was filed on January 3, 2025, and the patent was granted on January 27, 2026, with 20 claims.

Its first claim describes a system that receives a query, builds a results page whose first result leads to an organization's landing page, calculates "a landing page score for the first landing page" and, when the score exceeds a threshold, generates an updated results page with "a navigation link to an AI-generated page for the first organization". In other words: an AI-generated page for that organization, which the user reaches through a link in the results.

The description gives examples of what the score can be based on: conversion rate, bounce rate, click-through rate, or a qualitative factor, "page design quality or content quality". It also shows which way the threshold points: a landing page without a filter for products can exceed it, for instance, and the system can generate a page "for a query that may have poor targeting and/or a poor landing page". The link "can be included in a sponsored content item", and the generated page can include an AI chatbot and "a sitelink to a product detail page". As for the effect, the text says the system "can improve performance of content providers by improving performance metrics (e.g., total conversions, conversion rates)".

What it doesn't say: the word "spam" appears nowhere in the text, and loading speed is not among the score's criteria. The patent describes an AI-generated page for an organization, reached through a link in the results, in some embodiments from a sponsored result; it describes neither a spam filter nor a penalty for slow sites.

And one more thing: a patent doesn't show what Google runs. As the USPTO explains↗, a patent gives the right to "exclude others from making, using, offering for sale, or selling" the invention; it obliges no one to use it. We don't know whether or where Google applies this system, and anyone who says they know should show the source.

What to do on your site

• Read the policies, then your pages. The documentation says it directly: "Sites that see a change after a spam update should review our spam policies". The useful question for every group of pages produced at volume: what does the reader learn here that they can't find elsewhere?
• Improve or merge before you delete. For core updates, Google writes that "Deleting content is a last resort". Weak pages are rewritten with information of your own or merged, with a 301 redirect, into a better page. The steps are in our content pruning guide.
• No quick fixes. The same guide says: "Avoid doing 'quick fix' changes (like removing some page element because you heard it was bad for SEO)".
• Speed belongs to page experience. The spam policies have no rule about speed. Google recommends LCP "within the first 2.5 seconds of the page starting to load" and warns that good results in Search Console's Core Web Vitals report or third-party tools "doesn't guarantee that your pages will rank at the top of Google Search results". It is worth it for your readers.
• No special markup, no file for AI. Google's guide to its generative AI features in Search says "Structured data isn't required for generative AI search, and there's no special schema.org markup you need to add" and that, for Google Search, "you can ignore tactics like 'chunking' content, creating unnecessary AI text files (like llms.txt), or pursuing inauthentic mentions".
• Recovery takes months. "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." When the update concerns link spam, the benefit those links brought is lost for good: "Any potential ranking benefits generated by those links cannot be regained."

The test we apply to any claim of the "Google hit X" kind is in our article on SEO myths and the control test. The full framework is in our sustainable SEO guide.

We also declare our interest: we sell an article engine that works with AI models. That is exactly why we care what Google actually writes, not what circulates. If you want to see where your site stands against these policies, the technical audit is the place to start.

Sources

• Google Search Status Dashboard, March 2026 spam update↗ and March 2026 core update↗, read on October 29, 2026.
• Google Search Central, Google Search spam updates and your site↗, updated December 10, 2025.
• Google Search Central, Spam policies for Google web search↗, updated August 28, 2026.
• Google Search Central, A guide to Google Search ranking systems↗, updated December 10, 2025.
• Google Search Central, Google Search's core updates and your website↗, updated December 10, 2025.
• Google Search Central, Google Search's guidance on using generative AI content on your website↗, updated October 1, 2026, and Google Search's guidance about AI-generated content↗, February 8, 2023.
• Google Search Central, Optimizing your website for generative AI features on Google Search↗, updated July 10, 2026.
• Google Search Central, Understanding Core Web Vitals and Google search results↗, updated December 10, 2025, and Understanding page experience in Google Search results↗, updated September 22, 2026.
• Google Patents, US12536233B1, AI-generated content page tailored to a specific user↗, granted January 27, 2026, read in full on October 29, 2026.
• USPTO, Patent essentials↗, read on October 29, 2026.

Corrections and updates

• October 29, 2026. Article rewritten. The first version dated the volatility to March 8–22 from third-party ranking trackers and described a two-week rollout; the Search Status Dashboard records the update from March 24 to March 25. It presented application 20260089127 as a Google patent for detecting AI-written text; that number belongs to another company's application, about a chatbot that keeps the context of a conversation (its Google Patents record↗), and the detection mechanisms the first version listed are not in it. It also said the helpful content system was folded into the core algorithm in 2023; Google's ranking systems guide dates that to March 2024. Also removed: figures and claims without a source (recovery in one to three core update cycles, a 20% drop threshold, a list of page types the update supposedly filtered) and a quote that tied a patent to spam.