Google AI search opt out removes sites from AI features

Google AI search opt out is now a live Search Console control that removes a site from AI Overviews, AI Mode, and generative AI features in Google Discover without changing ordinary Google Search ranking.

Google Search Console Help says Google rolled out the Search generative AI control to all websites worldwide on August 31, 2026. Zyppy then used the control on September 1, 2026, opted back in on September 29, and published its test on Zyppy Signal on November 5. Search Engine Journal reported the test on November 9, writing that Cyrus Shepard, founder of Zyppy SEO, estimated AI Overviews and AI Mode accounted for about 21% of one site's Google search clicks, or 23% when he included a small estimated lift to regular listings.

How Google AI search opt out works

Google's control removes eligible site content from Search generative AI features, including links and grounding material, while leaving ordinary web search eligibility intact. Google Search Console Help says excluded sites will not receive traffic or impressions from AI Overviews, AI Mode, or covered Discover generative AI features. Google also says the control is not used as a ranking or inclusion signal for other parts of Search and does not affect AI training.

The Search generative AI control changes where Google may show and use a site in search results, not whether the site can rank in normal web listings. If a site is excluded, Google says its content cannot be visible to users in covered AI features, cannot be linked there, and cannot help ground those AI responses.

The timing is one place where the primary source and the test experience differ. Google says exclusion generally takes a few days, with exclusion within 1 to 2 days after the control goes live, although some content may take longer. Search Engine Journal reported that Shepard said Zyppy disappeared from AI features in less than an hour, and Zyppy Signal says AI-feature impressions fell by more than 99% after the opt-out.

What Zyppy measured after switching off

Zyppy measured the effect by comparing a 28-day AI-OFF period with a 28-day AI-ON period using Google Search Console data. Zyppy Signal says the test used Web performance data, including clicks, impressions, CTR, average position, queries, and landing pages, plus Google's Generative AI Features report for AI-specific impressions. Shepard described the work as a single experiment on a single site and told readers to treat it "more as a directional yardstick than Google gospel."

Shepard estimated AI clicks by comparing total clicks in the AI-OFF period with total clicks in the AI-ON period, while controlling for the same searches and similar ranking positions. Search Engine Journal says his comparison produced an estimated 0.9% click-through rate for Zyppy's AI Overview and AI Mode impressions, compared with around 0.4% across Google Search as a whole. Because AI features made up about 10% of the site's search impressions, Shepard estimated they produced about 21% of its search clicks.

Zyppy's result does not prove that every site should stay in Google AI features. It shows that one site with about 10% of impressions in AI Overviews and AI Mode estimated a higher AI-feature CTR than its overall Google Search CTR, then lost those measurable AI impressions when it used the opt-out.

The test also found a possible secondary effect that is harder to isolate. Search Engine Journal reported that Shepard saw a small CTR lift for regular organic listings when Zyppy also appeared in AI Overview results, with the largest effect on URLs that had the most AI visibility. Shepard called this a halo effect and said nobody knows for sure why it happens, though brand reinforcement is one possible explanation.

What Search Console still hides

Search Console still hides the click split that publishers need most, because clicks from AI Overviews and AI Mode are counted inside the standard web Performance report. Search Engine Journal says neither Google's generative AI performance documentation nor its Performance report documentation lists a way to separate clicks on links in AI features from other web clicks. That is why Shepard had to estimate AI clicks rather than read them directly.

The evidence is thin outside Zyppy's site, and the limits matter. Zyppy had done work to win AI citations, according to Search Engine Journal, which may make its AI click share higher than sites with weaker citation visibility. The test design controlled for the same queries and positions, but it was not a multi-site study and it does not show how ecommerce, local, news, or affiliate sites would perform.

What search teams should change

Search teams should measure AI visibility before excluding a site, because Google says an excluded site receives no impressions or traffic from the covered AI features. The specific check is the Generative AI Features report in Google Search Console, paired with the Web performance report for clicks, impressions, CTR, average position, queries, and landing pages. A defensible test needs the same query set and landing-page set across the before and after period.

For most publishers, the practical move is to treat the control as a test switch, not a default setting. If AI impressions are material, document the start date, the Search Console property, the affected URL patterns if any, the AI impressions trend, and the ordinary web ranking positions before changing the setting. If traffic falls after exclusion while normal rankings hold steady, the lost visits are consistent with AI feature removal, but Search Console will still require estimation until Google reports AI clicks directly.

What this means for AI traffic

The inference for search visibility is that AI citations can be a traffic source, not only a zero-click risk, but the value depends on query mix and citation share. Zyppy's site saw AI-feature impressions equal to about 10% of total search impressions and estimated those impressions had higher CTR than its Google Search average. That does not overturn publisher complaints about AI answers, but it does show why opting out before measurement can remove traffic that a site cannot yet see as a separate click line.

The paid media angle is less settled, but the pressure point is visible in Alphabet's search business. Search Engine Journal reported on November 5 that Google Search queries hit a five-year growth high while Search and other revenue grew 12% to $63.35 billion in the third quarter, slower than earlier 2026 quarters. The same report says Google chief business officer Philipp Schindler named AI Mode query growth as one cause because it "had an impact because of the lower monetization we were seeing in AI Mode."

For SEO and paid search teams, that creates a measurement problem on both sides of the SERP. Organic teams need to know whether AI citations send visits or lift ordinary listing CTR. Paid teams need to know when query growth moves into a format with lower ad yield, because impressions and clicks may not map to older search revenue patterns. Both conclusions are inferences from the reported evidence, not proof that AI Mode will keep lower monetization or that Zyppy's click share will generalize.

The next forcing date is December 3, 2026. Search Engine Journal says the UK Competition and Markets Authority gives Google until then to provide publishers with impressions, clicks, and click-through rate for AI search features in the UK, aggregated across features such as AI Overviews and AI Mode. Google then has three more months to roll out page-level controls for AI search features, which is the next trigger for publishers deciding whether exclusion should happen at site level or page level.