Google's Spam Policies and AI Search: What Changed in May 2026, and What the Rules Say About AI Overviews and AI Mode

Do Google's spam policies apply to AI Overviews and AI Mode?

Yes. Since 15 May 2026, Google's spam policies have defined spam as "techniques used to deceive users or manipulate our Search systems into featuring content prominently," and the examples now include "attempting to manipulate generative AI responses in Google Search" alongside manipulating rankings. Google's documentation for site owners describes two such features, AI Overviews and AI Mode, so tactics aimed at getting cited in them are judged by the same rules as tactics aimed at ranking.

Google recorded the edit in the changelog for its Search documentation under 15 May 2026: "Clarified that our spam policies also apply to generative AI responses in Google Search." The reason it gave was "To make it clear that the spam policies apply to all of Google Search, including generative AI responses." Barry Schwartz of Search Engine Roundtable reported the change that morning, after comparing the page with an archived copy from 13 May, and noted that Google's statements were on the changelog rather than in a blog post, with no Google employee quoted.

The policy page itself does not mention AI Overviews or AI Mode by name, and it lists no tactics specific to AI. What applies to AI answers is the existing list, which runs from cloaking, doorway abuse and hidden text to link spam, scraping and user-generated spam, plus the three policies Google added in March 2024: scaled content abuse, site reputation abuse and expired domain abuse. An earlier version of this article said Google named six specific practices in the update. Google's text names none.

How have Google's spam policies changed since 2024?

By accumulation rather than reversal. Google added three policies in March 2024, read one of them more strictly in November 2024, published guidance on AI-generated content in 2025, and in 2026 added a malicious-practices rule, a new use for spam reports, the generative AI clarification and a different enforcement regime in Europe.

The May clarification arrived on the same day as Google's new guide to optimizing for generative AI features, and the guide is where Google spelled out what it means in practice. On pages written to catch every phrasing of a question, it says: "While it might be tempting to create separate content for every possible variation of how people might search (for example, by focusing on other queries that people have asked, or fan-out queries), doing so primarily to manipulate rankings or generative AI responses in Google Search violates Google's scaled content abuse spam policy." On manufactured buzz, it says "seeking inauthentic 'mentions' across the web isn't as helpful as it might seem."

Why do AI Overviews and AI Mode follow the same spam rules?

Because they are built on the same index and ranking systems as the ordinary results. Google's guide says its generative AI features "are rooted in our core Search ranking and quality systems," and adds: "Our core ranking systems focus on high-quality content while other systems block spam; our generative AI features depend on both."

The eligibility rules make the link direct. To appear as a supporting link in AI Overviews or AI Mode, Google's page on AI features and your website says, a page must be "indexed and eligible to be shown in Google Search with a snippet," and "There are no additional technical requirements." The spam policies say sites that violate them "may rank lower in results or not appear in results at all." A page that spam systems have demoted or removed has lost its route into the AI answer as well.

The same page describes query fan-out, "issuing multiple related searches across subtopics and data sources," which means an AI answer can cite a page that ranks for one of those related searches rather than for the question a person typed. That mechanism is also the source of the temptation Google warned about in May: if an answer draws on dozens of sub-questions, the obvious shortcut is a page for each one. Our guide to how AI Overviews and AI Mode choose their sources covers fan-out and eligibility in detail, and the honest version of the work.

Is AI-generated content against Google's spam policies?

No, not in itself. Google's position since February 2023 has been that "Appropriate use of AI or automation is not against our guidelines." What breaks the rules is using automation, AI included, to produce content whose main purpose is manipulating rankings, which is the same test Google applies to content written by people.

That position comes from a February 2023 post by Danny Sullivan and Chris Nelson, under the heading "Rewarding high-quality content, however it is produced." It also advised accurate author bylines where readers would expect them, disclosure where someone might wonder how a piece was made, and observed that "Giving AI an author byline is probably not the best way" to tell readers that AI was involved.

In May 2025 Google turned this into a standing page, its guidance on using generative AI content. It says "Generative AI can be particularly useful when researching a topic, and to add structure to original content," and then draws the line: "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." It reminds publishers that "generative models don't retrieve facts, but predict a likely sequence of words based on their training data," and that the review for accuracy applies to titles, meta descriptions, structured data and image alt text as well as body copy. Merchants get one hard rule: AI-generated product images in Merchant Center must carry the IPTC DigitalSourceType value TrainedAlgorithmicMedia in their metadata.

The October 2026 update to that page points to the Search Quality Rater Guidelines on scaled content abuse (section 4.6.5) and on content made with "little to no effort, little to no originality, and little to no added value" (section 4.6.6). Google adds that raters' ratings "don't directly influence ranking"; the raters' job is "to help evaluate the performance of our various search ranking systems."

In practice, a team that uses AI to draft, then checks every claim, adds what only it knows and publishes pages it would publish anyway is on the right side of all of this. A team that uses AI to publish pages nobody asked for, at a volume no editor could read, is the case the policy was written for.

What counts as scaled content abuse?

Google defines it as "when many pages are generated for the primary purpose of manipulating search rankings and not helping users," typically "large amounts of unoriginal content that provides little to no value to users, no matter how it's created." The method of production is irrelevant. The purpose and the value are what count.

The policy arrived in March 2024, when Google's announcement said it would cover scaled content "whether automation, humans or a combination are involved." Google said it expected the March changes to cut low-quality, unoriginal content in results by 40%, and in a note dated 26 April 2024, after the rollout finished on 19 April, said people would "now see 45% less low-quality, unoriginal content in search results." The policy page gives these examples:

• "Using generative AI tools or other similar tools to generate many pages without adding value for users"
• Scraping feeds, search results or other content to generate many pages, including by synonymizing or translating it, "where little value is provided to users"
• "Stitching or combining content from different web pages without adding value"
• "Creating multiple sites with the intent of hiding the scaled nature of the content"
• "Creating many pages where the content makes little or no sense to a reader but contains search keywords"

The AI search version of this is easy to recognise: a page for every city in a service area with only the place name changed, a "best X for Y" page for every combination in a spreadsheet, or a page for each fan-out question an AI tool suggests. Google's May guide calls the last of these a violation when it is done to manipulate rankings or AI responses, and adds that it is "also an ineffective long-term strategy, as a high quantity of pages doesn't make a website higher quality or more relevant to users." One page that answers a set of related questions well serves both the reader and the fan-out.

What is site reputation abuse, and what changed in November 2024 and August 2026?

It is third-party content placed on an established site to borrow that site's standing in search. Google's policy, now headed "Site reputation policy," applies "where third-party content is published on a host site mainly because of that host's already-established ranking signals, which it has earned primarily from its first-party content," with the goal of making the content "rank better than it could otherwise on its own."

Google's examples of a violation are an educational site hosting third-party sponsored reviews of payday loans, and a medical site hosting a low-quality third-party advertising page about "best casinos." Its examples of what is not a violation include wire services and press release sites, syndicated news, forums and comment sections, columns and opinion pieces, advertorial content meant for readers rather than rankings, and affiliate links that are handled appropriately.

Enforcement began on 5 May 2024. In November 2024 Google's Chris Nelson closed the obvious loophole. After reviewing white-label services, licensing agreements, partial ownership and other arrangements, Google concluded that "no amount of first-party involvement alters the fundamental third-party nature of the content," and warned that "we don't simply take a site's claims about how the content was produced at face value." A publisher that oversees a partner's coupon or product-review section is still hosting third-party content.

The August 2026 change was about Europe. "Following discussion with the European Commission, we are adjusting our enforcement approach," the Search Quality team wrote on 28 August. From 30 August, a manual action under the policy still hits the affected section of a site in results shown outside the European Economic Area, but "For users inside the EEA, the impact of the manual action won't apply." Inside the EEA the section may instead be separated from the main domain in Google's systems and left to rank on its own merits over time. Google said it was concerned that "an overbroad application of the DMA could prevent us from addressing real threats," referring to the EU's Digital Markets Act.

For AI search, the policy closes off the most obvious shortcut: a vendor-written "best software for X" article placed on a strong publisher's domain in the hope that AI answers cite it. That arrangement fits Google's description whether the target is a ranking or an AI answer. A forum thread does not, though the policies separately list "Spammy posts on forum threads" as user-generated spam. Our guide to Reddit, Google and AI Overviews covers how a business can take part in forums without crossing that line.

What is expired domain abuse?

It is buying a lapsed domain to trade on its old reputation. Google's policy describes it as "where an expired domain name is purchased and repurposed primarily to manipulate search rankings by hosting content that provides little to no value to users." Its examples are affiliate content on a site that once belonged to a government agency, and casino content on a site that once belonged to an elementary school.

The test is the purpose and the value of what goes on the domain, so the policy is aimed at the use rather than the purchase. An old domain's standing lent to unrelated, low-value pages is the pattern it describes, and aiming those pages at AI answers instead of rankings earns no exemption.

Can you pay to get cited in AI Overviews or AI Mode?

Not through placements or links without running into one of the existing policies. Paying for links meant to influence Google is link spam, paying to place your content on a stronger site is what the site reputation policy describes, and buying an old domain for its reputation is expired domain abuse. Since May 2026, none of those becomes acceptable because the target is an AI answer rather than a ranking.

Google's policy defines link spam as "the practice of creating links to or from a site primarily for the purpose of manipulating search rankings," and its examples include "Exchanging money for links, or posts that contain links" and "Using automated programs or services to create links to your site." It also says "Google does understand that buying and selling links is a normal part of the economy of the web": paid links are fine as long as they are qualified with a nofollow or rel="sponsored" attribute.

Mentions are covered by the May guide. Google's AI features "can show what's being said about products and services across the web, including in blogs, videos, and forum discussions," it says, but "seeking inauthentic 'mentions' across the web isn't as helpful as it might seem." The reason is in the sentence about systems quoted above: the AI features depend on the same systems that block spam.

So when a vendor promises citations in AI Overviews or AI Mode, ask exactly what it will do. If the answer is paid placements, links or a network of sites, it is selling one of the three practices above with a new label. The legitimate routes are slower: pages worth citing, coverage a journalist chose to write, and participation in the forums your buyers read, under your own name. Whether any of it works shows up in how often assistants name you; our guide to measuring AI visibility as share of model explains how to sample that honestly.

How does Google enforce its spam policies, and how often?

With software first and people second. The policy page says Google detects violations "both through automated systems and, as needed, human review that can result in a manual action." Google also releases periodic spam updates, which it announces on its Search Status Dashboard.

Three of this year's four updates, in June, August and September, came after the May clarification. Google said nothing about what any of them targeted, so there is no basis for calling any of them an AI spam update.

The human side grew in 2026. On 14 April Google clarified that it "may use spam report submissions to take manual action against violations," and on 23 April clarified again "To address feedback we received about the change on using spam reports to take manual action." A report filed by anyone, a competitor included, can now lead to a manual action. The back button hijacking rule announced on 13 April took effect on 15 June, and Google's announcement said violations could bring manual actions or automated demotions. Sites that keep trying to work around the rules face a separate provision: Google "may take appropriate action which may include restricting or removing eligibility for some of our search features."

A manual action is reported to the site owner in Search Console, as Google's August post restates for site reputation actions. For a drop with no manual action behind it, the evidence is timing: a fall that lines up with a rollout window on the dashboard. Before blaming a spam update, compare the pages that fell with similar pages that did not, over the same weeks. Our guide to measuring a change against a control group explains the method, and the free before-and-after check for Search Console runs it on an export.

What should a site owner check now?

Anything on the site that exists mainly because a search engine, or an AI answer built on one, is watching. The checks below map to the policies above:

• Pages generated per city, product variant or suggested question. Each should be worth publishing if search did not exist; merge the thin ones into pages that answer the set.
• Third-party sections on your domain, such as coupons, partner reviews and sponsored guides. Know who writes them and why they are there. Since November 2024, your oversight does not change their status.
• Paid and sponsored links, in both directions. They need rel="sponsored" or rel="nofollow".
• AI-drafted pages. Check the facts, the title, the meta description, the structured data and the alt text, and disclose AI use where a reader would wonder.
• Scripts from ad and recommendation vendors that interfere with the back button.
• The Manual actions report in Search Console, after every spam update on the dashboard.
• Any vendor contract that promises AI citations. Find out what the vendor actually does.

For the crawling and indexing side of the same review, our technical SEO audit checklist covers what stops a page from being indexed or shown with a snippet, the two conditions for appearing in AI answers at all.

One question settles most cases faster than the policy text. If a tactic only makes sense because a search engine is watching, it fits the definition Google extended in May. If it would still be worth doing with search switched off, it does not.

Questions