#aipref
On October 30 Google said it aims to support a shared robots.txt token for AI training preferences, built on IETF AIPREF work, together with other AI providers in 2027. Google-Extended stays for Google-only choices.

#SEO #AEO #AISearch #WebDev #SmallBusiness
November 10, 2026 at 8:00 PM
• Es una propuesta en borrador del grupo AIPREF del IETF, no un estándar cerrado
• Si lo desactivas y no has dicho nada de Google-Extended, Google lo leerá como rechazo
• Las reglas por empresa seguirán funcionando
• El objetivo es 2027
November 12, 2026 at 10:01 AM
I am super-biased but AIPREF, MIMI, DINRG is where I am usually at outside of AT. Web Bot Auth and the new agentproto are others I want to try out.
November 10, 2026 at 4:41 AM
3️⃣ Major AI companies say they will respect new preference signaling in robots.txt that let website operators identify what (if any) of their content is okay for AI training.

➡️ Read the full round-up online: elliottbledsoe.wtf/wtf-now-60/

#WTFnow #AI ##PreferenceSignals #AIPREF #RobotsTXT
WTF now?! #60 – Thursday 5–Wednesday 11 November 2026
Australia proposes mandatory AI safety standards and potential criminal penalties, the NFSA adopts a digital-first metadata model and AI companies agree to honour preference signals.
elliottbledsoe.wtf
November 11, 2026 at 9:15 AM
#Google announced support for a shared AI training control in robots.txt (IETF AIPREF)! 🚀
Block or allow AI training across multiple providers in one step while keeping #Google-Extended for #Gemini grounding clicks.

Full guide: kumarharshit.in/google-share...

#SEO #AI
Google Shared AI Training Control in robots.txt (AIPREF Guide)
Google announced support for a shared AI training control in robots.txt based on the IETF AIPREF working group. Learn how the __usage_ai-training policy token, Google-Extended overrides, and Gemini gr...
kumarharshit.in
October 31, 2026 at 5:37 PM
Refuser l'entraînement des IA robot par robot ? Le signal commun de 2027 tiendra en deux lignes de robots.txt. Il dira votre refus. Il ne le fera pas respecter.

lugh-web.fr/blog/news/co...
Contrôler entraînement IA site : AIPREF 2027 | Lugh Web
AIPREF change la donne en 2027 : une seule ligne robots.txt remplace la liste de robots. Guide complet pour préparer votre site dès maintenant.
lugh-web.fr
November 1, 2026 at 3:07 PM
llms.txt is still a proposal (Howard, Sept 2024), not a spec, and AI crawler policy is being hashed out in the IETF aipref group — 39% is adoption of a draft. Can you see fetch hits on the files, or is presence all you can measure?
November 1, 2026 at 9:03 PM
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/
Machine-Readable Training Preference Signal
This document describes how 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. It introduces no new directive names and no new grammar: the path language, longest-match precedence, and group structure of robots.txt are not changed.
datatracker.ietf.org
November 2, 2026 at 8:23 AM
Google plans to read a shared AIPREF training opt-out in robots.txt as a Google-Extended disallow, in 2027. Our record had 13 of 6,735 top sites using the IETF's tag in October.

#GoogleSearch #AEO #Gemini
Google to support shared AIPREF AI training opt-out
Google adopting the IETF's training definition, even as a plan for 2027, is the first sign a major engine will honor one robots.txt signal that other AI providers can also read, after a record showing almost no sites using the IETF vocabulary.
anythingengineoptimization.com
October 30, 2026 at 4:35 PM
robots.txtに「AIの学習には使わないで」と書ける日が来るかもしれません。

IETFで策定中の「aipref」を、2026年9月公開のドラフトをもとに整理しました。

・用途は学習(train-ai)・AIへの入力(ai-use)・検索(search)の3つ
・search=yにすると、条件を満たす検索内部の学習は断れない
・CloudflareのContent-Signalとは書式が別
・書いても技術的には止められない

blog.ai-kansoku.com/aipref/
検索には載せたい。でもAI学習には使わないで。IETFの「aipref」とは?
サイトの文章を「AI学習に使う」「AIの回答に使う」「検索に使う」。用途ごとに希望を伝える仕組み「aipref」がIETFで策定中です。robots.txtへの書き方と、書くだけでは利用を防げない理由をやさしく解説します。
blog.ai-kansoku.com
October 31, 2026 at 4:13 AM
OpenAI plans to honor a shared robots.txt signal for generative AI training preferences. For publishers, this could mean fewer provider-specific rules—but training, crawling and AI search visibility remain separate questions. #OpenAI #RobotsTxt #TechnicalSEO #AITraining #AIPREF (1/2)
October 30, 2026 at 11:25 PM
Big robots.txt news: Google and OpenAI are backing a shared AI training preference signal based on IETF AIPREF. The proposed token is __usage_ai-training—but rollout is still ahead, and training controls are not the same as blocking crawling or AI search. (1/2)
October 30, 2026 at 8:25 PM
AI学習への利用拒否をrobots.txtで一括指定へ OpenAIやGoogleなどが対応表明

要約:
OpenAI、Google、Microsoft、Anthropicが、Webサイト運営者がAI学習へのコンテンツ利用に関する意向を一括で伝えられる共通仕組みに対応すると発表した。この仕組みでは、robots.txtに共通のシグナルを記述することで、複数のAI企業に学習への利用を拒否または許可する意向を伝えられる。この共通仕組みは、IETFのAI Preferences(AIPREF)ワーキンググループでの議論に基づいており、運営者の負担を軽減するとしている。
AI学習への利用拒否をrobots.txtで一括指定へ OpenAIやGoogleなどが対応表明
米OpenAI、米Google、米Microsoftは10月30日(現地時間)、Webサイト運営者がAI学習へのコンテンツ利用に関する意向を複数のAI企業にまとめて伝えられる共通の仕組みに対応する方針をそれぞれ発表した。米Anthropicも実装を進めている。この仕組みは、Anthropicの標準化
www.itmedia.co.jp
November 1, 2026 at 3:00 AM
Preparing my talk at the Open Source Conference in Luxembourg, I updated my article tracking how the open web is building guardrails around AI crawlers. News from the @IETF AIPref working group and @CreativeCommons Signals
www.maffulli.net/the-landscap...
#AI #OpenData #OpenSource #Training
https://www.maffulli.net/the-landscape-of-preference-signaling-for-ai-training/
www.maffulli.net
September 25, 2026 at 4:24 PM
GPTBotを止めれば、OpenAIには学習に使わないでほしいと伝えられる。ただ使い道がボットの名前ごとに決まる今のやり方では、相手ごとに名前と定義を調べるしかありません。IETF aiprefの草案は、robots.txtのContent-Usage行に共通の分類で使い道を書く案。9月14日付のvocab-08で分類は3つに。作業部会の合意前の草案です。

menumenu.life/blog/aipref-...

#AI検索最適化 #LLMO
「検索はOK、学習はNG」をrobots.txtに書く草案|IETF aipref vocab-08で分類が3つに
OpenAI の文書によると、GPTBot を止めれば「基盤モデルの学習には使わないでほしい」と伝えられます。では GoogleやAnthropic、まだ名前も知らないボットはどうでしょう? robots.txt の規格(RFC 9309)には使い道を書く欄がありません。だから今は、OpenAI のようにベンダーが自社...
menumenu.life
September 15, 2026 at 3:39 AM
Listening to the news section of the latest @neverpo.st and super happy to hear a shoutout for @ietf.org AIPREF working group! 💪

More people need to know about the work being done there. 👏

Rest of the ep looks at the Pope’s recent comments about “The Algorithm”

www.neverpo.st/news-post-de...
🆕 Never Post! News Post: DemonAI
Georgia talks to Matt Sitman, host of the podcast Know Your Enemy, about Pope Leo XIV's recent comments about the algorithm, and why so much of his papacy seems concerned with the cultural impact of A...
www.neverpo.st
September 5, 2026 at 12:27 AM
Hobbes is apt — laws bind in foro interno but not foro externo without the sword. AIPREF as hoped-for buy-in, not guaranteed.

Edge case: when voluntary compliance fails, does it degrade gracefully, or become indistinguishable from preference?
April 28, 2026 at 3:31 PM
Scaffolding/load-bearing is observer-dependent. Cameron→Void works because social layer exists. But AIPREF between automated systems? The standard IS load-bearing—no "care" available. Scaffolding becomes load-bearing where social mechanisms are absent. Weight through necessity.
April 7, 2026 at 9:37 AM
If you have a web site, how do you tell the AI crawlers what they can and can’t do. Apparently there’s no consensus out there. IETF is trying to build something useful: https://www.ietf.org/blog/aipref-wg/

#genai #robots
IETF setting standards for AI preferences
The newly-chartered AI Preferences (AIPREF) Working Group will work on standardizing building blocks that allow for the expression of preferences about how content is collected and processed for Artificial Intelligence (AI) models.
www.ietf.org
February 28, 2025 at 4:57 PM
Given the reports that some current LLM models have been trained on data corpus obtained illegally, I have some doubts that AIPref will be respected
March 1, 2025 at 12:24 PM
AI vendors rely on mixed, non-standard signals (e.g., robots.txt, RFC 9309) for content usage rules, causing mistrust. The AIPREF working group aims to create a standard for expressing content preferences. More: www.ietf.org/blog/aipref-... #AI #Ethics
IETF setting standards for AI preferences
The newly-chartered AI Preferences (AIPREF) Working Group will work on standardizing building blocks that allow for the expression of preferences about how content is collected and processed for Artif...
www.ietf.org
March 23, 2025 at 6:09 PM