← All articles

AI Training Preferences: Google and OpenAI Make a Move

Google and OpenAI have each published an announcement about how websites can express their AI training preferences. Here is what we know, what we don't know yet, and how a website owner in our region can get ready.

By Publicus Group · November 12, 2026 · 4 min read · Lire en français
Photo : Annie Spratt (Unsplash)

The essentials in 30 seconds

  • Google has published a post about a shared control that lets sites express their AI training preferences, and OpenAI, for its part, announces a simpler way to do it.
  • The two announcements appear to be related, but nothing in our sources confirms it; the technical details, timelines and possible effects are not specified here.
  • If you have a website, now is a good time to decide where you stand and to find out who on your end manages these settings.

The context

Artificial intelligence tools learn from large amounts of content, some of which comes from websites. For a small business, a municipality or an organization, the question is simple to ask: do we want the content of our website to be used to train AI models, yes or no?

Until now, many website owners didn't have a clear, uniform way to answer. A new development spotted by our monitoring touches on exactly this point: two major players in the tech world have each published an announcement about how websites can express their AI training preferences.

The topic is new. It directly concerns anyone who owns a website, whether it's a restaurant in Dieppe, a firm in Moncton or a public organization.

What's changing

What Google is announcing. On its blog for website owners (Search Central), Google has published a post about a shared control that lets sites express their AI training preferences.

What OpenAI is announcing. For its part, OpenAI announces a simpler way for sites to express those same preferences.

The link between the two. The two announcements appear to be related, meaning they could be part of the same mechanism or the same initiative. However, our monitoring list does not confirm this. It would therefore be unwise to assume they are one and the same standard before reading both texts.

What we don't know yet. Here is what the sources we have do not specify, and which we therefore don't want to invent:

  • the exact technical method for expressing a preference;
  • the effective date, if there is one;
  • the possible consequences, positive or negative, on a site's visibility;
  • how other AI companies will respond to these preferences.

These details may be found in the announcements themselves, which we encourage you to read directly.

What this means for your organization

It's first a decision, not a technical task. Before touching anything, leadership should have a position: can our content (text, photos, blog posts, service descriptions) be used to train AI models? Some organizations will see an advantage, others an intellectual property concern. There is no universal answer, and we don't claim to have one.

It mostly affects those who produce original content. A site that publishes analyses, guides or in-house photos has more to decide than a site with a few pages of contact information. That said, even a small site benefits from knowing what it allows.

You need to know who controls the site. In many small businesses, the site was built by an outside provider, and no one remembers who holds the access credentials. If that person is hard to reach, this development is a good reminder: you should be able to change your site's settings or, at the very least, know who to ask.

There's no need to panic, either. Nothing in the announcements we found points to an obligation or a deadline. A calm read, followed by a thoughtful decision, is enough for now.

What to do now

  1. Read both announcements: Google's post on AI training control and OpenAI's announcement on site preferences. Check for yourself whether they describe a common mechanism.
  2. Define your position at a leadership meeting: do you allow it, refuse it, or want to make exceptions depending on the type of content? Write down the decision and the reasoning behind it.
  3. Identify the person responsible for your website (employee, agency, freelancer) and confirm that you have access to the settings.
  4. Ask your provider the right questions: does this development apply to our site? How do we apply it? What effects should we expect? If the answer is vague, ask them to cite their official sources.
  5. Take stock of your sensitive content: documents, client photos, licensed or third-party text. These are the first items to protect if you choose to restrict.
  6. Schedule a follow-up in a few months: these announcements are recent, and the practical details may change. Add the topic to your digital monitoring.

If you're short on in-house resources, remember that you don't have to sort it all out alone. A conversation with the person who manages your site, or with a trusted digital advisor, is often enough to clarify the way forward.

Sources

Let’s talk about your project

Start your project in 10 minutes

Start a project