One-sentence summary
This FAQ explains what the October 30, 2026 robots.txt AI-training opt-out signal agreed by OpenAI, Google, Microsoft and Anthropic is, how to configure it, when it takes effect, and how it differs from Google-Extended.
Common Questions
Q1: What exactly is this robots.txt signal? A: An industry agreement announced October 30, 2026: a common policy label in robots.txt lets a site express opt-out from multiple AI vendors' training at once, discussed at the IETF AI-preferences working group.
Q2: Which vendors pledged respect? A: OpenAI, Google, Microsoft and Anthropic. Anthropic already adopted it; Google and Microsoft plan 2027; OpenAI in coming months. Meta supports the idea but has not committed.
Q3: Will it affect my pages being indexed? A: Per Microsoft, opting out of training (and anchoring) does not affect Bing indexing; Google-Extended only excludes Gemini training, not regular indexing.
Q4: How is it different from Google-Extended? A: Google-Extended is Google's single-vendor crawler control since 2023; the new signal aims to cover multiple vendors in one config and is still rolling out.
Q5: How should I configure it now? A: Keep existing rules like Google-Extended short term, watch the IETF label spec, and switch to the unified signal once OpenAI, Google and Microsoft fully support it.