IETF draft proposes robots.txt-based signal for AI training opt-outs
A new Internet-Draft lets website operators mark paths as off-limits (or open) to AI training using ordinary robots.txt syntax.
What to know
- The draft defines a way to signal 'no AI training' (or an opt-in) for specific paths using standard robots.txt syntax, requiring no new grammar.
- Legacy parsers that don't recognize the new policy-token user agent will ignore it and behave exactly as before, per the draft's design goals.
- It is an early-stage Internet-Draft (work in progress, valid six months) under the IETF AI Preferences Working Group, not an adopted standard.
- Public engagement so far is minimal: the Hacker News submission has a score of 1 and no comments.
Kelley Author of the Internet-DraftIETF AI Preferences Working Group Working group discussing the draft
How it unfolded 2 developments, newest first · click a bar or a number to jump articlesposts
-
2
Draft surfaces on Hacker News with minimal traction
User mehdizare posts the draft to Hacker News; as of the latest data it has a score of 1 and no comments, indicating the discussion has not yet developed.
“A service operator binds usage preferences (e.g., "no AI training") to paths by writing ordinary robots.txt groups keyed by a per-purpose policy token in the User-Agent line.”
— draft-kelley-aipref-training-preference-signal-00, IETF Internet-Draft text · source -
D
Finalmente uma forma simples de controlar o que se quer ou não para treinar a IA (ainda em draft). Agora basta é as empresas de IA seguirem realmente isso. 🤷♂️ https:// datatracker.ietf.org/doc/draft -kelley-aipref-training-preference-signal/
-
- 1 day quiet
-
1
IETF datatracker publishes AI-training preference draft
The Internet-Draft draft-kelley-aipref-training-preference-signal-00, 'Machine-Readable Training Preference Signal,' is listed on the IETF Datatracker, proposing a robots.txt-based mechanism for signaling AI training preferences.
-
first by IETF Datatracker, 12d ago
-
-
background
Draft details a backward-compatible robots.txt extension — The document specifies that preferences are expressed via a per-purpose policy token in the User-Agent line, introduce no new directive names or grammar, support path-specific bindings (e.g., disallowing training sitewide but allowing it under /press-kit/), and leave behavior unchanged for parsers that don't implement the mechanism.