Publishers are pushing back against Meta’s planned terms changes around AI scraping. The problem is that most of them still cannot afford to leave Meta’s platforms behind.
The contradiction is very clear: Publishers may dislike the idea of their journalism being crawled and used in AI products without compensation, but Meta remains valuable for much more than sending people back to their websites.
Meta’s value has moved beyond referral traffic
Facebook referral traffic has already fallen sharply for many publishers. One head of audience told Digiday that Facebook had dropped from more than a fifth of the publisher’s global traffic to a negligible share. Across the top 10 publisher websites tracked in the report, Facebook traffic was down 5% year over year.
That decline has not made Meta irrelevant. Instagram is still being used for brand awareness, community engagement and marketing. Meta’s platforms also support branded content, paid media, audience reach, conversion and subscription acquisition. Chartbeat data cited by Digiday shows Meta accounted for an average of 3.47% of global publisher traffic in July 2026 across more than 2,500 domains, but referral share does not capture all of the platform’s influence.
That is the uncomfortable shift. Meta can send fewer readers directly to a publisher and still shape how that publisher finds audiences, builds recognition and converts subscribers.
The exit threat is real, but still mostly theoretical
Publisher trade groups including the News Media Alliance, Digital Content Next and the Professional Publishers Association are pressing Meta to change the planned terms. The Guardian’s director of public policy, Tami Hoffman, has warned that some publishers could eventually leave Meta altogether. Other executives quoted by Digiday described the terms change as another example of Meta using its power to force an industry-wide choice.
But the publishers interviewed are not ready to abandon their social accounts. The situation resembles the long-running argument over Google’s crawlers: publishers can object to AI scraping while remaining dependent on the discovery and commercial systems around the platform.
The proposed answer from the U.K. publisher trade body the PPA is to separate these permissions. Publishers should be able to allow link previews or advertising while blocking AI training and product development, with protections such as robots.txt backed by binding terms rather than public assurances.
That would give publishers more control without requiring them to choose between protecting their content and remaining visible where audiences already spend time. Until that kind of separation exists, the threat to leave Meta will remain powerful rhetoric, but a difficult business decision.