Google has started giving some site owners a dedicated Search Console control for deciding whether their pages can appear in Google’s generative AI search features, including AI Overviews, AI Mode, and generative AI features in Discover. The change gives publishers a cleaner opt-out path than the blunt tools they have had until now, but it also forces a difficult traffic decision: a site that excludes itself from those AI surfaces gives up the links, impressions, and potential visits that come from them.
The control is rolling out first to a subset of website owners while Google tests it, and Google’s help documentation says the setting began being taken into account on June 17, 2026. It sits under Settings > Search generative AI in Search Console and offers three choices: include the site, exclude the site, or inherit the setting from a parent property.
For publishers, media companies, ecommerce sites, and independent site owners, the important point is that this is not the same as blocking Google Search. Google says the setting is not used as a ranking or inclusion signal for other parts of Search. A site can opt out of being used in the covered generative AI features while still remaining eligible for regular search results.
What the new control actually covers
Google’s Search generative AI control documentation lists three covered surfaces for now: AI Overviews, AI Mode, and generative AI features in Google Discover. Google says it expects that list to change as Search develops.
If a site owner chooses the exclude option, links to that site and content from that site are prevented from appearing to users inside those covered Search generative AI features. Google also says content crawled from the site will not be eligible as an input for generating an AI response or preview in those features. The company warns, however, that similar content from other sites can still appear.
The exclusion normally takes a few days to propagate. Google’s help page says content should generally be excluded within one to two days after the control goes live, though caching and broader propagation can make some removals take longer.
The setting also has an inheritance model that matters for larger sites. A child URL-prefix property can inherit the control from a parent domain property unless the owner configures the child property separately. That means publishers with subdomains, verticals, international editions, or section-level Search Console properties should check which property actually controls the content they care about.
The new reports are the other half of the decision
Google introduced the control alongside new Search Generative AI performance reports in Search Console. In its June 3 announcement, Google said the reports provide dedicated views of impressions from generative AI features in Search and Discover, while those impressions remain part of the overall performance report.
The reports show which pages appeared in generative AI features, where those appearances happened by country, what devices users were on for Search results, and how visibility changed over time. Google is also offering hourly, daily, weekly, and monthly granularity.
That reporting makes the opt-out less theoretical. Before turning the control off, a publisher can look for pages that already receive meaningful AI-search exposure, compare those impressions with ordinary search performance, and decide whether the visibility is useful enough to keep. A small publisher that gets little traffic from AI Overviews may view exclusion as a low-cost bargaining or content-control move. A site whose evergreen guides are regularly cited in AI Mode may have a harder call.
What it does not do
The control does not remove a site from Google Search. It also does not block crawling, prevent Google from understanding pages for core Search, or fully address AI training. Google’s documentation says the control does not affect AI training and points site owners to Google-Extended if they want to limit training of models used to generate responses in Search generative AI features.
That distinction is where many site owners can get tripped up. Search Console’s new switch governs participation in certain Google Search AI answer surfaces. Robots.txt, Google-Extended, noindex tags, snippets controls, and separate commercial licensing terms each handle different parts of crawling, indexing, display, and training. Treating one switch as a universal AI-content policy would be a mistake.
Google’s broader guidance for generative AI features also pushes back on special AI-search hacks. The company says its generative Search features rely on core Search ranking systems, retrieval-augmented generation, and query fan-out. It also says Google Search ignores llms.txt files for visibility or ranking, even if other AI services choose to use such files.
Why this is happening first in the UK
The timing is tied to regulatory pressure. The UK Competition and Markets Authority announced on June 3 that Google must give publishers more control over use of their content in AI search under the country’s digital markets competition regime. The CMA described the requirement as a world first and said publishers should be able to prevent their content from powering AI features in Search while preserving their ability to appear in traditional search results.
The CMA also said Google must make sure publisher content is properly attributed with clear links in AI-generated search results, and that publishers need controls over fine-tuning use cases as well. Google’s rollout is therefore both a product change and a policy response to a larger dispute over how AI search redistributes traffic, bargaining power, and attribution.
How site owners should approach the switch
The safest first step is measurement, not a reflexive opt-out. Site owners who have access should review the generative AI performance report, identify the pages that appear most often, and compare those pages with organic search, Discover, subscription, affiliate, ecommerce, or ad-revenue goals.
For a news publisher, the key question may be whether AI Overviews are summarizing reporting in ways that reduce clicks or weaken licensing leverage. For an ecommerce site, the question may be whether AI search visibility sends qualified traffic, supports product discovery, or simply answers comparison queries before a shopper reaches the site. For a help-center or tutorial site, the calculation may depend on whether brand visibility inside AI answers is valuable enough to offset fewer direct visits.
Large sites should also test the inheritance rules carefully. A domain-level exclusion may affect more sections than intended, while a section-level override may be useful for publishers that want news, reviews, commerce pages, or archives treated differently. Any change should be documented with the date, property, scope, and reason so traffic shifts can be interpreted later.
For now, Google’s new toggle is best understood as a specific control over covered AI search appearances, not a complete answer to AI scraping, training, publisher compensation, or search-market power. But it is still a meaningful change: site owners finally have a first-party way to separate regular Google Search visibility from participation in some of Google’s answer-generating search products.
Sources
- Google Search Console Help: Search generative AI control
- Google Search Central: Introducing Search Generative AI performance reports in Search Console
- Google Search Central: Optimizing your website for generative AI features on Google Search
- UK Competition and Markets Authority: CMA secures fairer deal for publishers and improves Google search services in UK
