Does Google-Extended Block AI Overviews?

Short answer: no

One robots.txt line can cut you out of an answer engine. Google-Extended is the line people worry about most, but it is the wrong line for AI Overviews.

Blocking Google-Extended does not block your pages from Google AI Overviews. AI Overviews ride on Google Search. The page has to be crawlable by Googlebot, indexed, and eligible to show a snippet in normal Search. There is no separate Google AI Overviews crawler to allow or block.

Google-Extended is a control token in robots.txt. It does not have its own HTTP crawl user-agent. Google says it controls whether content already crawled by Google may be used for future Gemini model training and some grounding uses. It does not affect inclusion in Google Search, and it is not a Google Search ranking signal.

The bot list that matters

Split the world into live answer visibility and training opt-outs. That removes most of the confusion.

• For Google AI Overviews, allow Googlebot. Blocking Googlebot can remove the page from Search and from AI Overviews.
• For ChatGPT search answers, allow OAI-SearchBot. Blocking it keeps you out of ChatGPT search answers, though navigational links may still appear.
• For Claude search visibility, allow Claude-SearchBot. Blocking it removes you from Claude search results that rely on that crawler.
• For Perplexity answer visibility, allow PerplexityBot. Perplexity also documents Perplexity-User for user-requested fetches. Its docs say that fetcher generally ignores robots.txt because a user asked for it.
• For Apple search and Apple Intelligence answer context, allow Applebot. Applebot-Extended is a control token, not a crawler.

By contrast, GPTBot, ClaudeBot, CCBot, Google-Extended, and Applebot-Extended are training, dataset, or opt-out controls. Google-Extended and Applebot-Extended are robots-only control tokens, with no separate crawler to look for in logs. Blocking them is a content-use choice. It does not, by itself, remove your pages from live AI-search visibility.

What to change in robots.txt

If you want visibility in AI answers, do the boring crawl checks first.

• Do not disallow Googlebot if you care about Google Search or AI Overviews.
• Do not disallow OAI-SearchBot, Claude-SearchBot, PerplexityBot, or Applebot if you want those systems to reach your pages.
• Use GPTBot, ClaudeBot, CCBot, Google-Extended, and Applebot-Extended when your goal is training or reuse control.
• Keep key copy in server-rendered HTML. Googlebot documents Search rendering. For other AI crawlers, client-side-only content is an undocumented risk, so do not make them execute JavaScript to find the answer.
• Make sure your CDN, WAF, geo rules, and login walls allow the same access your robots.txt file allows.

Robots.txt is a stated policy. It is useful, and major crawlers publish how they read it, but it is not proof of what a request did. Some user-triggered or reported bots, including Perplexity-User and Bytespider, have been reported to ignore robots.txt in some settings. Treat logs as evidence only after you verify IP ranges or reverse DNS where the vendor publishes them.

If you want a quick read on which answer engines can fetch your site, use the free AI visibility checker. It is the right test for this question.

Do not mix this up with email deliverability

Crawlability decides whether a page can be found. Email authentication decides whether a message can be trusted.

For email, Gmail and Outlook.com care about SPF, DKIM, DMARC, MX, reputation, complaint rate, engagement, and blocklists. A clean SEO robots.txt file will not save a domain with broken mail records.

• SPF is a TXT record that lists who can send for your domain. Publish one SPF record, keep it under the 10 DNS-lookup limit from RFC 7208, and avoid +all. Use ~all while testing, then -all when you know every sender is covered.
• DKIM signs mail with a private key and publishes the public key at a selector, such as selector._domainkey.example.com. If the provider is not signing, the DNS record alone does nothing.
• DMARC checks alignment. The visible From domain must align with passing SPF or DKIM. Start with p=none and a rua report address, read the reports, then move to p=quarantine or p=reject when legitimate mail passes.
• MX records tell the world where to deliver inbound mail. Bad MX does not usually explain outbound spam placement, but it can break replies and domain trust checks.
• Blocklists matter most after abuse, compromise, or bad list buying. Fix authentication and sending behavior before chasing delisting forms.

If your mail is drifting, run a free domain scorecard at InboxRadar. For aggregate DMARC XML, the free DMARC report reader turns RUA reports into plain senders and failure reasons.