Shared AI Training Control for Websites: What Google and OpenAI Announced

What Google announced

John Mueller, writing for Google Search Central on October 30, 2026, said Google plans to support a shared AI training control in robots.txt. Instead of maintaining a separate instruction for every AI developer, site owners would use a shared "policy token" in robots.txt to set broad preferences across participating providers.

Google said this builds on Google-Extended, the AI training control it has offered since 2023. When a website uses the shared token to opt out of training, and Google-Extended is not explicitly specified, Google plans to interpret that as if the publisher had disallowed Google-Extended. Other providers, Google wrote, can handle the same token similarly. Company-specific controls will continue to work alongside the joint signal, so a site can set a broad choice and still tailor instructions to an individual developer such as Google.

Google also stated that websites that specifically allow Google-Extended are eligible to appear in responses and receive clicks via grounding in Gemini Apps. That is a separate decision from whether a page appears in Google Search features such as AI Overviews. For Search appearance, see our article on AI Overviews and answer engine optimization.

Google said the proposal reflects publisher, civil society, and regulator feedback calling for industry-wide training controls, and that it is consistent with the EU Code of Practice for general-purpose AI model providers under the EU AI Act, which recognizes robots.txt as a primary machine-readable method for publishers to express preferences. Google aims to support the new control in 2027 and said it will provide updates as work continues through the IETF working group.

What OpenAI announced

The same day, OpenAI published "Making it easier for websites to express AI training preferences." OpenAI said it will honor a common signal in robots.txt for the use of website content in foundational generative AI training. OpenAI described this as a voluntary effort that reduces the work of maintaining separate rules for every provider.

OpenAI wrote that it aims to implement the signal in the coming months and will share another update with implementation details when it does. The new signal will not replace the controls OpenAI already provides through robots.txt. OpenAI said its existing crawler controls will continue to apply and be respected. Like Google, OpenAI pointed to the IETF AI Preferences working group and encouraged other providers to honor the signal.

The two timelines are not the same. Google is pointing to 2027. OpenAI is pointing to the coming months. Site owners should watch each provider's follow-up posts rather than assuming one date covers both.

What to do now

1. Decide what you want, in plain language, before you edit files. Training a model on your public pages is not the same thing as allowing Google to crawl the site for Search, and it is not the same thing as allowing grounding in Gemini Apps. Write down which of those you want before anyone changes robots.txt.
2. Review the controls that already work. Google-Extended has been available since 2023. OpenAI says its current crawler controls remain in force. If your WordPress site or other CMS already ships a robots.txt file, look at what it actually contains today rather than assuming it is blank or complete.
3. Do not invent a shared token syntax. Google described a policy token but did not, in the October 30 post, publish a final robots.txt snippet for site owners to copy. Wait for each provider's implementation notes. A guessed directive can fail silently.
4. Keep Search visibility separate from training preferences. If you block Googlebot in robots.txt, you affect Search crawling, including AI features that rely on that crawl. Google documents separate preview controls such as nosnippet and noindex for what Search can show, and Google-Extended for some other Google AI systems.
5. Revisit the file when implementations ship. Google said it will provide updates on its way to 2027 support. OpenAI said it will share details when it implements the shared signal. Put a reminder on the calendar rather than treating the October 30 posts as a finished technical standard.

Why this matters for NTx clients

Website owners have been asking a simple question: can AI systems train on our site? Until now the practical answer has been a list of vendor-specific crawler names. A shared token, if providers actually honor it, would make that answer easier to implement. It would not, by itself, decide whether your pages show up when someone asks ChatGPT or Gemini a question, and it would not replace backups, updates, or access control.

On an aging WordPress site, robots.txt is easy to overlook. A plugin, a security tool, or a previous developer may have written rules that no one on the team remembers. Reviewing that file is part of ordinary WordPress Security hygiene, and it now has an AI-training consequence as well as a search-crawling one.

If you want help separating website controls from a broader AI plan, that is the kind of question we take up in an AI Readiness Consultation.

FAQ

Is the shared AI training control live today?

Not as a finished, copy-and-paste standard for every provider. Google said it plans to support the shared robots.txt control and aims to do so in 2027. OpenAI said it will honor the common signal and aims to implement it in the coming months. Existing crawler-specific controls remain the tools that work now.

What is Google-Extended?

Google-Extended is the AI training control Google says it has offered since 2023. In the October 30, 2026 announcement, Google said company-specific controls such as Google-Extended will continue to function alongside the new joint signal. Google also said sites that specifically allow Google-Extended are eligible to appear in responses and receive clicks via grounding in Gemini Apps.

If I opt out of AI training, will I disappear from Google Search?

Not automatically. Google presented the shared token as an AI training preference. Search crawling is still governed by Googlebot and related Search controls. Blocking Googlebot is a different decision from disallowing Google-Extended. Google documents separate preview controls for how snippets appear in Search, including AI features.

Should I add a robots.txt line today that I found on a blog?

Not unless that line comes from a current official provider document. Google's October 30 post describes a policy token but does not give a final snippet to paste. A guessed directive may do nothing, or it may conflict with rules you already rely on.

Does this apply to WordPress sites?

Yes, if the site is publicly crawlable. WordPress sites typically have a robots.txt file at the root. Review what it contains, keep Search crawling and AI-training choices distinct, and change the file only after you know which providers you want to allow or disallow.

Sources

1. John Mueller, "Supporting a shared AI training control for websites," Google Search Central Blog, October 30, 2026. https://developers.google.com/search/blog/2026/10/ai-training-control
2. OpenAI, "Making it easier for websites to express AI training preferences," October 30, 2026. https://openai.com/index/helping-websites-express-ai-training-preferences
3. Google Search Central, "AI features and your website" (Search crawling, preview controls, and Google-Extended). https://developers.google.com/search/docs/appearance/ai-features