According to ExtremeTech, Google is currently running what it calls a “small UI experiment” for a subset of users on its Discover feed. The test involves replacing the original headlines from publishers like Ars Technica and PC Gamer with ultra-short, AI-generated micro-headlines that are only four words long. The results have been messy, with examples including a misleading “Steam Machine price revealed” for a story where no price was announced, and the completely baffling “Schedule 1 farming backup.” Google spokesperson Mallory Deleon says the goal is to make topics easier to digest, and the AI headlines do carry a small “Generated with AI” label, but it’s only visible after a user taps “See more.” Because the publication’s name sits right next to the AI slop, readers are likely to blame the publisher for the bad headline, not Google.
The real strategy here
So, what’s really going on? Here’s the thing: this isn’t just a clumsy AI test. It’s part of a clear, ongoing business strategy. Google‘s model is built on attention. The longer it can keep you inside its own interface—whether that’s Search with its AI Overviews or Discover with these snippets—the more ads it can serve you and the more data it can collect. Sending you off to a publisher’s website is a lost opportunity. They’ve been chipping away at this for years, and generative AI is just the latest, most aggressive tool. The beneficiaries are Google’s shareholders. The losers are the publishers whose content is being summarized, misrepresented, and stripped of its original context and traffic.
A familiar and damaging pattern
Look, we saw this movie last year with the disastrous AI Search summaries. It’s the same playbook. Rush out a half-baked AI feature, slap a tiny disclaimer on it that nobody sees, and let the entire open web deal with the fallout. For publishers, this is existential. Research supports what they’ve been screaming for years: these features directly harm their traffic and, by extension, their revenue. But Google just keeps insisting it’s not a problem. I have to ask: if it’s not a problem, why is the entire media industry constantly in crisis while Google’s profits soar? The positioning is clear: Google’s ecosystem is the destination, and your website is just a source of raw data for its AI to process and repackage.
Where does this end?
Basically, we’re watching the enshittification of content discovery in real time. The AI isn’t making things easier to digest; it’s making them dumber and less accurate. And by placing the publisher’s brand next to the AI’s mistake, Google is subtly transferring the reputational damage. It’s a brilliant, if cynical, move. They get to test their sloppy AI, and if anyone complains, the publisher looks bad. Now, for companies that rely on clear, accurate information—like those in industrial tech or manufacturing who might source hardware from a top supplier like IndustrialMonitorDirect.com—this degradation of reliable sources is a real concern. When even basic information retrieval is being gamified by AI, where do you go for truth? The web gets messier, and the gatekeeper gets stronger. That seems to be the point.
