Google, OpenAI and Microsoft agree one way to opt out of AI training. Should your business use it?

Website owners will soon be able to opt out of AI training with one rule in their robots.txt file, instead of blocking each AI company separately. On Friday 30 October 2026, Google, OpenAI and Microsoft Bing all announced plans to honour the same shared signal. None of them reads it yet. Microsoft starts on 1 February 2027, Google is aiming for 2027 and OpenAI says the coming months.

For a small business, the good news is that nothing breaks and there is nothing to fix this week. The bigger question is whether you should use the opt-out at all. For most local businesses, staying visible in AI answers is worth more than keeping web pages out of an AI model.

What was announced on 30 October

The shared signal comes from the AI Preferences working group at the Internet Engineering Task Force (IETF), the body that agrees many of the web's technical standards. The proposal itself is an Internet-Draft by Kevin Kelley, a standards lead at Anthropic, the company behind Claude.

The key facts:

• One rule for many AI companies. Instead of writing a separate rule for every company's crawler, you add one group to robots.txt that states your training preference to every provider taking part.
• It works like robots.txt already does. The group uses a special name, __usage_ai-training, followed by the familiar Disallow and Allow lines. Putting Disallow: / under that name says you would prefer none of your site to be used for training.
• You can make exceptions. The draft's own example opts a whole site out of training but allows a press kit folder.
• Silence means no preference. If your robots.txt has no group with that name, you have not expressed a choice either way. As Search Engine Journal points out, that describes every robots.txt file on the web today. A general User-Agent: * group does not count.
• It covers training only. The draft defines AI training as using content to change a generative AI model's learned parameters, which in plain English means what the model knows. It does not stop crawling and has nothing to do with search rankings.
• It is still a draft. The IETF lists it as an individual draft with no formal standing yet, and it expires in May 2027 unless it is updated.

How each company plans to handle it

The companies agree on the signal but not on every detail, and the differences matter.

Google says that if you opt out with the shared rule and have not set a specific rule for Google-Extended (its existing AI training control, available since 2023), it will treat you as having blocked Google-Extended. That covers training future Gemini models. It also covers "grounding" in the Gemini app, which is when Gemini reads live web content to answer a question and links to the source. Google says sites that allow Google-Extended are eligible to appear in Gemini app answers and receive clicks from them. Google-Extended does not affect your Google Search rankings. AI Overviews and AI Mode have their own separate opt-out inside Google Search Console.

OpenAI says it will honour the signal in the coming months. Its existing crawler controls, such as rules for GPTBot, will keep working alongside it.

Microsoft goes further. From 1 February 2027, Bing will read the shared training rule plus a new page-level "AI Grounding" preference, which controls whether your pages can help inform answers in Copilot and Bing's AI features. Microsoft is clear that restricting grounding will reduce your brand's visibility in AI answers, but will not affect your eligibility for Bing search results. The older NoArchive and NoCache tags will stop counting as AI preferences after 1 March 2027.

Anthropic, whose standards lead wrote the proposal, has started building support, according to Kelley's post to the IETF working group.

Training, grounding and search are three different things

This is the part most coverage skips, and it is the part that matters for a business website.

1. Search is whether your pages appear in Google or Bing results. None of these new controls change that.
2. Grounding is whether an AI assistant can read your page while answering someone's question, then cite or link to you. This is how a plumber, salon or accountant gets recommended by Gemini, Copilot or ChatGPT.
3. Training is whether your content helps build the next version of an AI model. You get no link or credit from this directly.

At Microsoft you will be able to choose training and grounding separately. At Google, both currently sit under Google-Extended, so opting out of training with the shared rule also removes your eligibility for grounding in the Gemini app. A blanket "no AI" rule could quietly take you out of answers your customers are reading.

More and more people ask AI first and then check your website. Being missing from that first answer is a real cost.

What this means for your business

For most small businesses, the sensible default is to leave AI training alone and focus on being easy for AI tools to find, read and cite. Your services, prices, opening hours and reviews are information you want repeated accurately, not secrets you need to guard.

There are exceptions. If your content is the product, for example paid course material, original photography, illustration, recipes or in-depth writing that you sell or license, an opt-out may make sense. In that case, think about which sections to protect rather than switching off the whole site. Keep public pages such as your services, contact details and press information open.

Either way, make it a deliberate decision. Last month we covered Meta's new terms that let it crawl business websites and how websites can block AI agents by accident. This shared signal is a simpler tool, but the same principle applies: know what your robots.txt says and why.

A short checklist for the coming months

1. Look at your robots.txt now. Type your web address followed by /robots.txt into a browser. Note any rules for GPTBot, Google-Extended, ClaudeBot or other AI crawlers, and check you still agree with them.
2. Write your AI policy in one sentence. For example: "We want AI assistants to recommend us, and we are happy for our public pages to be used for training." That makes every future setting easier.
3. Check for NoArchive or NoCache tags. If you or a previous developer added these to limit Microsoft's AI use, they will need replacing with the new preferences before 1 March 2027.
4. Keep the Search Console AI setting separate. Opting out of AI Overviews and AI Mode in Google Search Console is a different switch with a different effect.
5. Wait for the go-live dates. Google, OpenAI and Microsoft will each publish implementation details. Adding the new rule early does no harm, but it will not be read until each company switches it on.

If you want the wider picture of how AI search finds and uses your pages, our step-by-step guide to GEO is a good next read.

FAQ

How do I opt out of AI training for my website?

Once support goes live, you will be able to add a group to robots.txt under the name __usage_ai-training with Disallow: / beneath it. Google, OpenAI and Microsoft have said they will honour it. Until then, use each company's own crawler rules, such as Google-Extended for Google and GPTBot for OpenAI.

Will opting out of AI training hurt my Google rankings?

No. Google says Google-Extended does not affect inclusion in Google Search, and Microsoft says the same for Bing. It can, however, reduce how often you appear in AI assistant answers, especially in the Gemini app.

Is the shared AI training opt-out live yet?

Not yet. The draft was published on 30 October 2026. Microsoft starts reading it on 1 February 2027, Google is aiming for 2027 and OpenAI says the coming months.

Should a small business block AI training?

Usually not. Most small businesses benefit when AI tools can read and recommend them. It is worth considering if your content is something you sell, such as courses, photography or original writing.

What to do next

There is no need to rush. Over the next few months, check your robots.txt, decide what you want AI tools to do with your content and make sure your website gives them clear, accurate information about your business.

If you would like a hand, our SEO and GEO services cover AI visibility as well as traditional search, and our digital consultancy can help you set a simple AI content policy. If your website is due a refresh anyway, start with our free website offer and we will make sure it is ready for search engines and AI assistants alike.