Home/News/Google Plans Shared AI Training Control in robots.txt, Targeting 2027

SEO & GEO

Google Plans Shared AI Training Control in robots.txt, Targeting 2027

2026-10-30 Google Search Central Blog SEO & GEO
Google Plans Shared AI Training Control in robots.txt, Targeting 2027
示意图(AI 生成,非现场图片)

On October 30, 2026, Google said on the Search Central blog that it plans to support a shared AI training control in robots.txt, letting a site set its AI training preference across multiple providers with one policy token. Google aims to support it, along with other AI providers, in 2027.

Key facts

  • When: not live yet; Google targets 2027
  • Who is affected: store owners who want to limit AI training on product copy, blog posts and images, and those who care about being cited in Gemini
  • Action: nothing for now. Existing Google-Extended rules keep working; revisit robots.txt once the final syntax is published

What happened

Until now you had to write separate rules per AI developer; Google's is Google-Extended, offered since 2023. The new approach builds on the IETF AIPREF working group. If a site opts out of training with the shared token and doesn't explicitly specify Google-Extended, Google plans to treat it as if Google-Extended were disallowed, and other providers can do the same. Company-specific controls keep working alongside it. Google also notes that sites that explicitly allow Google-Extended are eligible to appear in responses and get clicks via grounding in Gemini Apps. This is still a draft proposal, and the rest of the AIPREF vocabulary is under discussion.

What this means for independent store owners

There is very little to do today: this is a plan, not a shipped feature, and your current robots.txt behaves exactly as before. The real point is the trade-off. Opting out with one shared token closes the door on every participating provider at once, while Google says only sites allowing Google-Extended are eligible for Gemini citations and clicks. If you rely on content and AI-search visibility to get traffic, blocking training entirely and being cited in AI answers may pull in different directions, so don't treat it as all-or-nothing. Two things worth doing now: check whether your robots.txt already has Google-Extended rules, and decide whether you are comfortable with your product copy and blog being used for training. Wait for the 2027 syntax and the list of supporting providers before choosing between the shared token and per-provider rules. The details are still thin, so no rush to edit anything.

Someone should be watching these changes for you

I run AI automation for independent stores — content, SEO/GEO, product listing, social, and keeping up with policy shifts like this one.

See how it works