One-sentence summary
This standard overview covers the IETF AI-preferences working-group draft on a common robots.txt AI-training opt-out label, tied to the October 30, 2026 OpenAI/Google/Microsoft/Anthropic agreement.
Standard Overview
Discussed by the IETF AI-preferences working group, the draft aims to define a common policy label in robots.txt so a site can express preferences for AI training, crawling and citation across vendors in one rule. It remains a draft, not yet an RFC.
Key Provisions
| Item | Content |
|---|---|
| Status | IETF working-group draft, under discussion |
| Core mechanism | Common policy label in robots.txt |
| Behaviors covered | AI training, crawling, citation preferences |
| Tied agreement | Oct 30, 2026 four vendors (OpenAI/Google/Microsoft/Anthropic) |
| Rollout | Anthropic live; Google/Microsoft 2027; OpenAI in months |
| Meta | Supports but not committed |
| Search indexing | Does not affect regular indexing such as Bing |
