Open or Closed: The Decision You Didn’t Know You Were Making
I’ve spent the last four months analysing 5,125 publisher domains across 99 countries. I wanted to know what publishers are actually doing about AI — and whether any of it looks like a decision.
Three numbers stood out:
21% block all AI crawlers. 47% allow all. 28% have set nothing at all.
Those look like positions. They aren’t. For most, it’s a CMS default from a different era — a robots.txt written before AI was a category, inherited from a previous platform, never revisited. The Nordic blocking rate of 67% isn’t a Nordic stance on AI. It’s a Nordic CMS configuration.
There’s a sharper number underneath. The global average AI Act readiness score is 4.4 out of 100. For a publisher facing the EU AI Act’s transparency provisions, that isn’t “slightly behind.” That’s running governance on accident.
I’ve watched publishers reach for the binary: block to protect the IP, or open up and hope to get cited. After four months in the data, I think both framings miss what’s actually happening.
Blocking protects your content from being used in training. It does not protect your authority from being replaced. When someone asks an AI assistant about your beat, the model reaches for whoever is available. If that isn’t you, it’s an aggregator, a content farm, or Reddit.
You can be open and still be invisible. A site that can be crawled but has no structured data, no named authorship, no editorial declaration — the AI sees “a page with text,” not “the local authority on crime in this city.”
The publishers who’ll navigate this well aren’t the ones who got the binary right. They’re the ones who decided deliberately, and then built the infrastructure to back the decision.
— What is our principle on AI: block, license, or open on our terms?
— Do our technical settings actually reflect that position?
— Where is this written down as an institutional decision, not a CMS checkbox?
Default is not neutral. It is a position — with consequences.