According to TechCrunch, Google announced a new pilot program on Wednesday where it’s testing AI-powered article overviews on participating publications’ Google News pages. The commercial partnership includes publishers like Der Spiegel, El País, The Guardian, The Washington Post, and The Times of India. Google will make direct payments to these publishers to compensate for a potential drop in click-through traffic. The AI summaries will only appear on those specific publishers’ Google News hubs, not in general Search or other parts of News. This follows a similar move in July where Google added AI summaries to its Discover feed. The company also announced the global rollout of its “Preferred Sources” feature for English users, with all languages coming early next year.
The Traffic Trade-Off
Here’s the thing: this is a fascinating, and frankly desperate, attempt to square a circle. Publishers have been terrified that AI overviews will destroy their website traffic, which is their lifeblood for ads and subscriptions. Google‘s answer? Pay them off, at least for this experiment. It’s a direct financial acknowledgment of the harm its own features could cause. The logic is that if users get a good AI summary right on Google News, they might be more “engaged” with the platform itself, even if they click less. But is “engagement” on Google’s property as valuable as a visit to the publisher’s own site? For the publishers in this pilot, the check from Google probably makes the gamble worth it. For everyone else, it’s a worrying precedent.
Walled Gardens and Preferred Bubbles
And let’s talk about that “Preferred Sources” feature going global. On one hand, it’s great you can prioritize your favorite, maybe subscribed, news outlets. But combined with AI summaries that might satisfy your curiosity without a click, aren’t we building the ultimate, algorithmically-assisted filter bubble? You tell Google what you like, and its AI then summarizes content from those sources for you. Where’s the serendipity? Where’s the exposure to a challenging viewpoint? Google says it’s working with “emerging voices,” but the features it’s building seem designed to let users comfortably narrow their focus, not broaden it.
The Broader AI News Landscape
This isn’t just about text. Google’s also testing audio briefings from these summaries. So now you can listen to a robot read a robot’s summary of human journalism. The push into real-time info with partners like The Associated Press for Gemini is another front in the same war: making Google’s apps (Search, News, Gemini) the destination, not the gateway. The “contextual introductions” for links are a nice touch, basically the AI telling you why you should bother clicking. It all feels like a managed retreat from the open web. The old deal was Google sends traffic, publishers create content. The new deal seems to be: publishers help train and legitimize Google’s AI, and Google throws them some cash while it builds a walled garden of summarized information. I think the big question is, will users actually prefer this? Or will they miss the depth and voice of a full article, even if it takes an extra click?
