🔍 Read the full analysis: How To Set An AI Training Opt-Out Across Major Platforms on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
OpenAI, Google, Microsoft and Anthropic have announced plans or support for a shared robots.txt signal that tells participating companies not to use a publisher’s content to train generative AI models. The signal is voluntary, does not block crawling or AI answers, and companies have set different implementation dates; Meta has voiced support without committing to honor it.
OpenAI, Google, Microsoft and Anthropic said on October 30 they will support a shared instruction in websites’ robots.txt files for publishers who do not want their content used to train generative AI models. The proposed signal could replace the need to maintain separate training opt-outs for each participating company, but it is voluntary, does not cover every use of web content and is not yet an official internet standard.
The proposed rule uses the user-agent name __usage_ai-training with Disallow: /. It is described in an individual Internet-Draft titled “Machine-Readable Training Preference Signal,” published on October 30 by Kevin Kelley, an Anthropic standards lead. The draft is associated with the IETF’s AI Preferences working group, but does not have formal standing in the IETF standards process.
The companies have announced different schedules. Anthropic says it has started implementing support, while OpenAI says it aims to do so “in the coming months.” Google’s documentation points to early 2027. Microsoft says it will honor the shared training preference from February 1, 2027, as it crawls sites. Those statements describe company plans and commitments, not a guarantee that every service will interpret the signal identically.
Microsoft also announced a separate page-level preference for AI Grounding, covering web content used to inform answers in Bing, Copilot and Microsoft 365. Microsoft says the new preferences will take effect February 1, 2027, and that its NoArchive and NoCache tags will stop counting as training or grounding preferences after March 1, 2027. The company says those choices do not guarantee inclusion, traffic or payment.
Infographic 01 · Who backs it, and when
One line in robots.txt, four companies, four timelines.
A shared signal against AI training. Only Anthropic says it has started; the others name dates or months. The standard itself is still a draft.
- 30 Oct 2026Draft published, four companies announce supportdoneIndividual Internet-Draft by Kevin Kelley (Anthropic)
- 30 Oct 2026Anthropic: “has already started implementing support”startedIts standards lead on the IETF list: “It’ll take us some time to add support across our services.”
- Coming monthsOpenAI aims to implement itplanExisting GPTBot and OAI-SearchBot rules keep applying
- 16 + 19 Nov 2026Working group meets at IETF 127, San FranciscoopenOn 21 Oct a chair said no consensus has been declared yet
- 1 Feb 2027Microsoft honors the new preferences as it crawlsplanTraining in robots.txt or per page; grounding per page
- 1 Mar 2027Bing stops reading NoArchive / NoCache as AI preferencesdeadlinePublish the new settings before this date
- 2027Google plans support; its docs say “early 2027”planA “no” maps to Google-Extended unless you set that token yourself
- 3 May 2027Draft version 00 expires unless updatedInternet-Drafts are valid for at most six months
Binding? No, voluntary
- OpenAI calls it a “voluntary effort”.
- The vocabulary draft does not “ensure that preferences are followed”.
- The draft “has no formal standing in the IETF standards process”.
- Private agreements may override it, the draft says.
Who else?
- Meta supports a common signal, but did not say whether it will honor it.
- Apple, Amazon, xAI, Perplexity: no statement in the sources reviewed for this article.
- The draft’s acknowledgments list contributors “for identification only”, not supporters.
Sources: IETF draft, 30 Oct 2026 · Anthropic on the IETF list · AIPREF minutes, 21 Oct 2026 · Google, 30 Oct 2026 · Google crawler docs · OpenAI, 30 Oct 2026 · Microsoft Bing, Oct 2026 · Meta (Andy Stone on X). Company plans are company statements.
Powered by Thorsten Meyer AI
For publishers, one shared instruction could make it simpler to communicate a preference about AI model training across several major providers. Until support is in place, however, site operators may still need company-specific controls. The source reports that existing rules for GPTBot, Google-Extended and ClaudeBot remain applicable; Microsoft also says its preferences do not change eligibility for Bing search, while Google says Google-Extended does not affect Google Search.
The distinction between training and AI-generated answers matters. The shared signal is framed as a preference against training, not a general ban on crawling or on using live web content to answer questions. Microsoft’s separate grounding control illustrates that these uses may require separate choices. Publishers should not assume that opting out of training also prevents their pages from appearing in search or being used to inform an AI response.
The announcement also does not establish a right to compensation or remove material from models that have already been trained. It gives publishers a common way to express a preference, but the practical effect depends on whether each service adopts and follows it.
Popular right now
De'Longhi 15-Bar Pump Espresso Machine with Milk Frother & Temperature Control
Breville Bambino Espresso Machine, Brushed Stainless Steel (BES450BSS)
Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation
As an affiliate, we earn on qualifying purchases.
How the Opt-Out Took Shape
The shared signal is built around work in the IETF AI Preferences working group, which is developing ways for websites to express machine-readable preferences. The draft’s acknowledgments name participants connected with technology companies, publishers and other organizations. That broad list does not itself mean those organizations have adopted the signal.
According to the source article, Search Engine Roundtable first reported the development, followed by coverage from The Verge and AFP as published by The Straits Times. The article says it reviewed company announcements, the Internet-Draft and the working group’s public records. It also discloses that the draft was written by an Anthropic standards lead, that AI tools including Claude assisted with the article, and that the publisher belongs to a network of websites affected by AI companies’ use of online content.
At a working-group meeting on October 21, one of its chairs said, “We haven’t declared consensus on anything yet.” That is consistent with the draft’s current status: the proposed syntax is public, but it has not become an IETF standard.
““has already started implementing support for this draft””
— Anthropic, in its announcement about the draft
Adoption and Coverage Remain Open
Whether every company will honor the signal remains unsettled. Meta spokesperson Andy Stone said the company strongly supports publishers’ ability to express preferences and called a common signal an important step, but did not commit Meta to following it. The source article reports finding no statement from Apple, Amazon, xAI or Perplexity in the materials it reviewed; that does not establish what those companies may decide later.
The proposal is voluntary and not yet an official standard. The exact timing and scope of support may also differ across services. Anthropic says it will take time to add support across its services, and Google, OpenAI and Microsoft have described future implementation plans. It is not yet clear how consistently the signal will be applied in practice or how sites can verify compliance.
The signal’s scope is limited to training preference. It does not block a crawler from accessing a site, undo training already completed, automatically prevent content from informing AI answers, or create a payment arrangement. Publishers will need to distinguish those separate uses when setting their policies.
Watch for Service Rollouts
Publishers can review their existing robots.txt directives and decide separately whether to restrict training, AI-answer grounding and search access. The source recommends adding the shared training rule if a publisher wants to express that preference, while retaining company-specific directives until providers have implemented the shared signal. This is a practical precaution, not a guarantee of how every provider will respond.
Site operators using Microsoft’s NoArchive or NoCache tags should track the company’s stated dates: the new preferences are scheduled to take effect on February 1, 2027, and those older tags will stop counting as training or grounding preferences after March 1, 2027. Companies’ documentation and implementation announcements will be the clearest way to check whether their plans have changed.
The next broader milestone is the IETF working group’s consideration of the draft. Whether it gains consensus, is revised or becomes a formal standard remains unknown. Publishers and users should also watch for explicit adoption statements from Meta and other AI companies before treating the signal as broadly accepted.
Key Questions
What does the shared AI training opt-out signal say?
The proposed robots.txt rule uses User-Agent: __usage_ai-training and Disallow: / to express that a publisher does not want participating companies to use its content for generative AI training.
Does the signal block AI companies from crawling a website?
No. The source describes it as a training preference, not a general block on crawling. It also does not automatically stop web content from being used to inform AI answers.
When will the companies support the signal?
Anthropic says it has started implementation; OpenAI aims for the coming months; Google’s documentation points to early 2027; and Microsoft says it will honor the signal from February 1, 2027. These are company-specific timelines.
Has Meta committed to honoring the opt-out?
No commitment was reported. Meta spokesperson Andy Stone expressed support for publishers’ ability to state preferences, but did not say whether Meta will follow this signal.
Does this setting control training on my AI chat conversations?
No. The proposed instruction applies to website content and crawlers. It is not an account setting for conversations with an AI assistant; chat-data preferences are handled through each service’s account controls.
Source: ThorstenMeyerAI.com
