This question causes real anxiety, and it usually arrives wrapped in a worry: I want to stop Google using my content to train its AI, but I am terrified that the moment I add Google-Extended to my robots.txt, my organic rankings will tank. It is a sensible fear born of how much robots.txt can break. In this case, though, the fear is misplaced, and Google has said so in plain language. Here is exactly what Google-Extended does, what blocking it costs and does not cost, and how to decide.
The short answer
No. Blocking Google-Extended does not lower your organic rankings, and it does not remove you from Google Search at all. Google states this directly in its own documentation. Google-Extended is a separate control from the Googlebot crawler that indexes your site for Search. It governs only whether your content is used to train future Gemini models and to ground answers in Gemini apps and the Vertex AI API. Blocking it changes none of your Search visibility, and it does not even affect AI Overviews, which run on the regular Search index. So you can opt out of AI training without any SEO penalty. The only thing you give up is participation in Gemini’s training and grounding, which is a real but different tradeoff.
What Google actually says
You do not have to trust a blogger on this, because Google put it in writing. In its crawler documentation, Google states plainly that Google-Extended does not impact a site’s inclusion in Google Search nor is it used as a ranking signal in Google Search. That is about as unambiguous as Google ever gets. The documentation describes Google-Extended as a control over using content to train future Gemini models that power Gemini apps and the Vertex AI API, and for grounding, meaning supplying index content to improve factuality, in those products. Nothing in that touches indexing or ranking.
What Google-Extended actually controls
To see why there is no ranking effect, understand what Google-Extended is and is not. It is not a crawler that fetches your pages for Search. The crawler that does that remains Googlebot, and it operates entirely separately. Google-Extended is a usage control, a signal that decouples whether your already-crawled content can be used for Gemini training and grounding from its normal use in Search. When you disallow Google-Extended, you are telling Google not to use your content for those AI-training purposes. You are not telling Googlebot to stop crawling, and you are not removing yourself from the Search index. The two are deliberately independent so publishers can make this exact choice without sacrificing search visibility.
Google-Extended versus Googlebot
The clearest way to hold this straight is side by side.
| Control | What it governs | Affects Search rankings? | Affects AI Overviews? |
|---|---|---|---|
| Googlebot | Crawling and indexing for Google Search | Yes, this is your Search visibility | Yes, AI Overviews use the Search index |
| Google-Extended | Using content to train Gemini and ground Gemini apps and Vertex AI | No, per Google | No, AI Overviews do not depend on it |
Read the bottom row twice, because it dissolves the fear. Blocking Google-Extended affects neither your rankings nor your appearance in AI Overviews. If you block Googlebot, that is a different and drastic action that does remove you from Search, but that is not what Google-Extended does.
Does blocking it remove you from AI Overviews?
No, and this surprises people. AI Overviews are built on Google’s regular Search index, the one Googlebot populates, not on Google-Extended training data. So a page can be blocked from Google-Extended and still be cited in an AI Overview, because that citation comes through the Search index. This is the same separation in action: Google-Extended governs Gemini training and grounding, while AI Overviews are a Search feature. If your goal was to avoid AI use of your content entirely, note that AI Overviews are not controlled by Google-Extended at all, they follow your normal Search and snippet settings instead.
So what do you actually give up?
Blocking Google-Extended is not free of consequences, it just has no SEO consequence. What you give up is twofold. First, your content stops contributing to training future Gemini models. Second, and more practically, your content can no longer be used for grounding in Gemini apps, which means Gemini is less likely to draw on and surface your content when users ask it questions. Since Gemini is one of the major AI assistants, opting out of grounding can reduce your visibility specifically within Gemini’s app experience over time. That is the real tradeoff: not rankings, but presence in one growing AI assistant.
The real tradeoff: a fast-growing channel
This is where the decision gets interesting, because AI assistants are not a fringe channel anymore. Adobe Analytics reported that traffic to US retail sites from generative AI sources jumped roughly 1,200 percent year over year by early 2025, and during the 2025 holiday season Adobe found AI-driven traffic up several hundred percent with AI referrals converting markedly better than other sources. Those figures are retail-specific, but they illustrate a broad reality: discovery through AI assistants is scaling fast. So opting out of Gemini grounding, while costless for SEO, does mean stepping back from a channel that is growing quickly. That is the genuine thing to weigh, not a ranking risk that does not exist.
Should you block it?
Decide on values and strategy, not on a rankings fear that has been put to rest. Block Google-Extended if you have a strong position against your content being used to train AI models, if you operate under licensing or rights constraints, or if you are pursuing content-licensing deals and want to preserve leverage. Allow it if you want maximum presence across AI assistants, including being grounded and surfaced in Gemini, and you are comfortable contributing to training. Many publishers reasonably allow it precisely because AI grounding is a visibility channel they do not want to forfeit. Either choice is legitimate. The point is that you make it for the right reasons, knowing your Search rankings are not on the line.
How to block it correctly
If you decide to block, it is simple and surgical. In your robots.txt, add a directive targeting the Google-Extended user agent and disallowing it, while leaving Googlebot fully allowed. The two are independent tokens, so disallowing Google-Extended does not touch Googlebot’s access. The common mistake is overbroad rules that accidentally also block Googlebot, which would genuinely harm your Search visibility, so keep the Google-Extended rule specific to that user agent and verify that Googlebot remains allowed. After deploying, confirm in your logs that Googlebot still crawls normally. This is the one place where a careless robots.txt edit could hurt you, not because of Google-Extended itself, but because of an accidental Googlebot block sitting next to it.
A worked example
A publisher wanted to opt out of AI training on principle but had frozen for months, convinced it would crater their traffic. Once they read Google’s own statement that Google-Extended is not a ranking signal and does not affect Search inclusion, they added a clean, Google-Extended-specific disallow, left Googlebot fully allowed, and watched their logs. Googlebot kept crawling, their rankings did not move, and they continued appearing in AI Overviews because those run on the Search index. The only change was that their content stopped feeding Gemini training and grounding, which was exactly the outcome they wanted. The months of hesitation had been based on a risk that did not exist.
Common misconceptions
A few related myths cause the panic. The first is that Google-Extended is a ranking signal, it is not, per Google. The second is that blocking it removes you from AI Overviews, it does not, since those use the Search index. The third is that it stops Googlebot crawling, it does not, the crawler is separate. The fourth is that blocking it is risk-free in every sense, when in fact the real cost is reduced presence in Gemini grounding, a strategic tradeoff rather than an SEO one. Hold the accurate picture: no ranking risk, a genuine AI-channel tradeoff, and one robots.txt pitfall to avoid (accidentally blocking Googlebot).
How this fits the bigger crawler picture
Google-Extended is one piece of the broader question of which AI crawlers and uses to allow. The same careful, per-agent thinking applies to other engines, for instance how Perplexity handles crawling and robots.txt, which I cover in can Perplexity AI actually crawl your website. And if your underlying goal is AI visibility rather than restriction, the levers are content and presence, detailed in Google AI Overview ranking factors, while the related myth that you need special markup is addressed in do you need specific schema markup for AI Overviews. The thread across all of them is to separate what a control actually does from what people fear it does.
How to think about your AI visibility either way
Whether or not you block Google-Extended, the question worth more of your attention is which AI-search queries matter in your niche and whether you show up for them. I use SQSEO for that, fanning one seed keyword into the question-level prompts that trigger AI answers, for free, so you can see where you have a citable angle. Blocking Google-Extended is a one-time policy decision. Being genuinely useful and present for the questions your buyers ask is the ongoing work, and it pays off across Search, AI Overviews, and the assistants you do choose to participate in.
A note on monitoring after the change
Even though there is no ranking risk from Google-Extended itself, treat any robots.txt edit as something to verify, because the file is easy to get subtly wrong. After deploying, do three quick checks: confirm in Search Console that Googlebot is still crawling and your pages remain indexed, watch your impressions and clicks for a week to confirm nothing moved, and re-read the rule to be certain the disallow targets only the Google-Extended user agent. If all three look normal, which they will when the rule is scoped correctly, you have done exactly what you intended, opted out of AI training with zero search cost. The discipline here is not fear of Google-Extended, it is good robots.txt hygiene.
The bottom line
No, blocking Google-Extended does not lower your organic rankings. Google states outright that it is not a ranking signal and does not affect Search inclusion, and it does not remove you from AI Overviews either, since those run on the Search index. What blocking it actually does is stop your content from training Gemini and being grounded in Gemini apps, which is a real strategic tradeoff given how fast AI assistants are growing, but it carries no SEO penalty. So make the decision on your stance toward AI training and your visibility goals, configure it carefully so you do not accidentally block Googlebot, and stop worrying about a rankings hit that Google has confirmed will not happen.
Frequently asked questions
Does blocking Google-Extended lower your organic rankings?
No. Google states in its documentation that Google-Extended does not impact a site’s inclusion in Google Search and is not used as a ranking signal. It is separate from Googlebot, the crawler that indexes your site, so blocking it does not affect your Search visibility at all.
Does blocking Google-Extended remove me from AI Overviews?
No. AI Overviews are built on Google’s regular Search index, which Googlebot populates, not on Google-Extended training data. So you can block Google-Extended and still appear in AI Overviews, which instead follow your normal Search and snippet settings.
What does blocking Google-Extended actually do then?
It stops your content from being used to train future Gemini models and from being used for grounding in Gemini apps and the Vertex AI API. The practical effect is reduced participation in Gemini’s training and app experience, not any change to Search rankings or AI Overviews.
Should I block Google-Extended?
It depends on your values and strategy, not on rankings. Block it if you object to your content training AI or have licensing reasons. Allow it if you want maximum presence across AI assistants, including being surfaced in Gemini, since AI is a fast-growing discovery channel. Either way, configure it to target only Google-Extended and leave Googlebot allowed.
