Can you stop Google using your website to train AI?
You want customers to find your website, but you may not want its articles used to train AI models. Google-Extended lets you limit certain Google AI uses without affecting inclusion or ranking in Google Search. There is a tradeoff: the same control also limits some uses of your content in Gemini answers.
What is changing, and when
TLDR AI’s November 2 edition pointed us to Google’s October 30 announcement of a shared way to say whether your content may be used to train AI models. The idea is to let website owners state a broad choice once in robots.txt, the file that gives automated visitors instructions, instead of maintaining separate instructions about training AI models for every provider.
Google aims to support the shared control in 2027. The linked proposal is an Internet-Draft, a work in progress rather than a finished standard. It is not a universal switch you can rely on today. Existing provider-specific controls remain relevant.
Newsletter discovery: TLDR AI, November 2Google’s announcement and planned timingIETF proposal: work in progress
Search and Gemini are different destinations
Google-Extended is an existing instruction in robots.txt. It covers training future Gemini models and grounding in Gemini Apps and certain Vertex AI services. Grounding means supplying web content to a model when it answers a question.
Google says this control does not affect inclusion or ranking in Google Search. If you want your material used in Gemini answers, weigh that opportunity before restricting Google-Extended.
AI Overviews and AI Mode are part of Google Search. Google documents their controls separately, using Googlebot access and page-level controls for snippets or indexing. Blocking Google-Extended does not remove your content from those Search features.
Google-Extended: affected products and Search distinctionGoogle Search: controls for AI Overviews and AI Mode
Choose the outcome before choosing the setting
Give your maintainer a business goal, not a request to “block AI”. Use this comparison to make that goal precise.
| Your goal | What to ask your maintainer | Tradeoff to accept |
|---|---|---|
| Limit Google’s use for Gemini training | Review Google-Extended rules. | The same control also covers the documented Gemini grounding uses. |
| Limit what appears in AI answers within Google Search | Review snippet and indexing controls for the relevant pages. | These are Search controls; changes can also affect ordinary search presentation or eligibility. |
| State whether providers may use your content to train AI models | Track the shared proposal and each provider’s adoption. | Google targets 2027. Do not assume all providers already honor it. |
Current Google Search controlsGoogle’s proposed treatment of the shared preference
Who should act now, and who can wait
If your business has decided to restrict training use of its public writing, ask your maintainer to review the existing controls now. You do not need to wait for a shared standard to make that preference clear.
If your priority is discovery in both Search and Gemini, pause before approving a broad restriction. First decide whether the training objection outweighs the Gemini uses you would also limit. Eligibility for an answer is not a promise of mentions, clicks or enquiries.
If you are simply curious about the new shared control, keep your current setup and revisit it when Google confirms support. Google plans to treat a shared instruction against using content to train AI models as a Google-Extended restriction when no explicit Google-Extended instruction is present. An existing explicit choice therefore deserves attention during any future change.
A useful brief for your website maintainer
Write down which public pages matter, whether Google Search discovery should stay available, and whether the covered Gemini uses are welcome. Ask the maintainer to explain which proposed rule affects each goal before changing anything.
Keep a copy of the current settings and the agreed purpose. After a change, ask for confirmation of the live rules and the relevant Search checks. You should be able to explain what changed and why.
Sources
Make your website’s priorities clear.
Share your public website URL and explain where you want customers to find it. forvertz can help define the website work around that goal.
Discuss your websiteUpdated