Google Extends Spam Enforcement to AI Overviews as Small Businesses Face New SEO Risks

Google, whose search engine remains central to parent company Alphabet‘s advertising business, updated its spam policy documentation in May 2026 to make one point clearer: attempts to manipulate AI Overviews and AI Mode will be treated like attempts to manipulate traditional search results.

The update extends an enforcement push that began with a major March 2024 policy overhaul. That earlier change introduced categories including scaled content abuse, site reputation abuse, and expired domain abuse. Google said at the time that the changes were designed to reduce low-quality, unhelpful content in search results by about 40%.

The latest change does not make AI-authored content automatically spam. Instead, it brings Google‘s generative search surfaces under the same policy logic already applied to web results. For small business operators, the question is no longer just whether their pages rank. It is also whether their content is trusted enough to appear in the AI-driven parts of search.

Google’s AI spam rules focus on manipulation, not AI use

Google‘s public position on AI-generated content has been broadly consistent since at least February 2023. The company says it evaluates content based on quality and usefulness, not simply on the tool used to produce it. The violation begins when automation, including AI, is used with the primary purpose of manipulating rankings.

The March 2024 update renamed the older category of spammy auto-generated content as scaled content abuse. That category covers the mass production of pages, whether written by AI or by human content farms, when the main purpose is to rank rather than help users.

Google also added two other categories. Site reputation abuse applies when a trusted domain hosts low-quality third-party or sponsored content to exploit its authority. Expired domain abuse targets the purchase of previously authoritative domains mainly to rank low-quality content under inherited trust.

“Appropriate use of AI or automation is not against our guidelines. What is against our guidelines is using AI or automation to generate content primarily to manipulate search rankings.”

– Google, Search Central documentation, February 2023

That statement defines the issue as intent and primary purpose, not authorship. It does not, however, explain how Google‘s systems determine intent at scale or how accurately those systems separate thin automated content from useful AI-assisted work.

The May 15, 2026, refresh added language confirming that manipulation of AI Overviews and AI Mode is covered by the same spam framework. Penalties can include ranking demotion or removal from Google Search. The commercial stakes are clear: the credibility of Google‘s AI answers depends on those surfaces not being overwhelmed by the same volume-based tactics that weakened parts of traditional search.

Enforcement evidence is visible, but the 40% target is difficult to verify

Google‘s claim that the March 2024 changes were designed to reduce unhelpful content by about 40% should be treated carefully. It was a company-stated goal, not an independently audited result. Wired and other outlets reported on the crackdown, including its focus on AI-generated clickbait and mass-produced articles, but there is no peer-reviewed or third-party audited public measurement showing exactly how much low-quality content was removed.

There is still clear evidence that enforcement happened. SEO practitioners and outlets tracking the update reported de-indexing and visibility losses for sites described as spammy, thin, or heavily dependent on scaled publishing workflows. Those observations matter, but they also come with limitations. Much of the available data comes from the SEO community rather than from Google‘s own enforcement disclosures.

Analysts, including Lily Ray of Amsive and Michael King of iPullRank, have argued that the practical target is not AI-assisted production as a category. The bigger risk is thin, scaled, low-value text that provides little beyond what a user could get from a generic AI prompt.

That distinction matters because SEO remains structurally important for small business revenue generation. A local contractor using AI to draft and refine a detailed FAQ about its services is operating very differently from a publisher pushing out hundreds of near-identical pages to capture long-tail searches. Google‘s policy recognizes that difference in principle, but the company has not published false-positive rates or accuracy benchmarks for the systems that enforce it.

King has also noted that even compliant sites can experience short-term ranking volatility after major updates while Google tunes its detection systems. That caveat reflects real uncertainty around automated classifiers and edge cases, although SEO firms also have a business interest in presenting the policy environment as complex enough to require professional support.

Small businesses face more risk when AI content is thin, repetitive, or lightly reviewed

The policy burden does not fall evenly. Large companies often have SEO teams, content strategists, legal reviewers, and compliance processes that can audit content portfolios and adjust quickly. Small businesses, especially sole proprietors and teams without in-house marketing staff, usually do not.

That resource gap matters because many small operators now rely on general-purpose AI writing tools such as ChatGPT, Jasper, Google Gemini, and similar platforms to create content at a volume they could not produce manually.

Survey data points in the same direction: small business owners are increasingly using AI tools to manage content production, marketing copy, and customer communication across multiple parts of the business. That can improve efficiency, but it can also create the exact pattern scaled content abuse is meant to catch when pages are published with minimal review or little original value.

The gray area is where many small businesses may feel exposed. A regional law firm that uses Gemini to draft dozens of location-specific practice pages may be trying to help local users, but those pages can still look repetitive and thin if they share the same structure and offer little unique detail. Google‘s public documentation does not explain exactly how those cases are resolved.

The expansion to AI Overviews and AI Mode adds another layer. Google has promoted generative search and related tools to businesses as discovery and engagement opportunities, including Google Gemini AI features for small businesses. Operators may now need to think not only about rankings, but also about whether their content is being cited, ignored, or suppressed in AI-generated answers.

Small businesses can reduce exposure by tightening AI content workflows

• Audit existing pages against the scaled content abuse definition – Review pages that are structurally similar, thin on original information, or hard to distinguish from generic AI output. The key question is not whether AI was used. It is whether the page provides unique value that helps a user more than a generic answer would. Weak pages should be consolidated, rewritten, or removed before enforcement forces the issue.
• Document the editorial process behind AI-assisted content – Google distinguishes between AI used to support human judgment and AI used to replace it at scale. Businesses should keep records of editing, fact-checking, and subject-matter review for AI-drafted pages. This will not guarantee protection, but it can support a reconsideration request through Google Search Console if content is flagged.
• Review sponsored and third-party content for site reputation abuse – Sites that host affiliate reviews, sponsored articles, or third-party content outside their core subject area should reassess those sections. A home improvement site hosting unrelated payday loan content, for example, risks enforcement that could affect more than the offending page.
• Check whether expired domains are being used to rank thin content – Businesses that have acquired domains for local keywords, older brands, or redirects should confirm those domains are not being used mainly to rank low-value content. Google‘s expired domain abuse policy applies whether the tactic was intentional or simply an opportunistic business decision.
• Keep Search Console access current and monitored – Google Search Console sends manual action notices when a human reviewer determines that part or all of a site violates spam policy. Those notices are different from algorithmic ranking drops and require a reconsideration request. Small businesses should make sure alerts go to someone who will act quickly.
• Separate AI-assisted drafting from automated publishing – Using ChatGPT, Gemini, or Jasper to create a first draft that a human substantially revises, verifies, and enriches is very different from publishing hundreds of minimally reviewed pages. Google‘s policy does not define a precise threshold, but the difference between assisted drafting and automated publishing is the right starting point for evaluating risk.
• Track AI Overview visibility separately from traditional rankings – Because the May 2026 update extended spam enforcement to AI Overviews and AI Mode, a page may perform acceptably in traditional results while still being excluded from generative surfaces. Tools such as Semrush and Ahrefs now offer AI Overview tracking, which can give operators earlier visibility into changes.

Search Console, policy updates, and AI Overview tracking will show where enforcement is moving

• Google Search Console manual action reports – The Security & Manual Actions section of Search Console is the clearest site-level signal that spam enforcement has moved beyond algorithmic suppression. Notices citing scaled content abuse or site reputation abuse would be especially important for small operators.
• Google Search Quality Rater Guidelines updates – Google‘s rater guidelines offer one of the closest public views into how the company explains quality to human evaluators. Changes to AI-generated content guidance, E-E-A-T expectations, or definitions of lowest-quality pages could indicate where enforcement is heading.
• Third-party AI Overview tracking from Semrush and Ahrefs – Semrush and Ahrefs both sell SEO tools, so their data should be read with that commercial context in mind. Still, category-level changes in AI Overview visibility over the months following the May 2026 update could help show which types of content are gaining or losing exposure.
• Enforcement reporting from trade and independent outlets – Wired, Search Engine Land, and Search Engine Roundtable regularly track update rollouts, de-indexing reports, and recovery cases. Clusters of small business sites reporting sudden traffic losses after confirmed updates would be an early sign of meaningful SMB impact.
• Google spam policy revision history – Google‘s spam policy page includes update history. Revisions to the definitions of scaled content abuse, site reputation abuse, expired domain abuse, or AI surface manipulation would show where enforcement authority is expanding before the impact is visible in rankings.

Google has now made clear that the same spam rules apply across traditional search and its AI-driven search surfaces. What remains less clear is how those rules will affect small operators that use AI to keep up with larger competitors but lack the staff to audit every page like an enterprise publisher. That uncertainty is the practical risk for small businesses: not that AI use is banned, but that thin, scaled, lightly reviewed content can now create exposure across more of Google Search than before.