Publishers are building walls. Faced with AI crawlers harvesting their archives to train models that will never send a reader back, they are locking gates, throttling bots, and putting more of the site behind logins and paywalls. This is a rational defense. It also has a second effect that gets less attention.

Every page you wall off is a page that no longer carries open-web display ads. The defensive move and the supply contraction are the same move. You cannot seal the content from scrapers without also sealing it from the ad-supported open internet that scrapers were riding on.

Defending the content shrinks the commons

The open web worked as a bargain: publishers gave away access, and advertising paid the bill. That bargain assumed the audience was human and the impressions were worth something. AI harvesting breaks the assumption. When a large share of your incoming traffic is machines extracting value and returning none, the open door stops making sense.

• Fewer open pages means fewer legitimate impressions in the auction.
• A thinner open-web supply pushes spend toward walled gardens.
• The walled gardens were already winning, and this hands them more.

So the story is not just AI versus publishers. It is that self-defense in an adversarial environment quietly enclosures the commons. Each individual wall is justified. The aggregate is an open web that is smaller, more concentrated, and harder for anyone new to enter.

The uncomfortable question is whether an ad-supported open web can survive an era where a growing share of its visitors are machines that were never going to convert. Publishers are answering it one gate at a time.