In early December, I brought you the news that Google has begun changing Verge headlines, and people of our rivals, with AI clickbait nonsense in its content material feed. Google appeared to be backing away from the experiment, however now tells The Verge that its AI headlines in Google Uncover are a characteristic, one which “performs nicely for person satisfaction.” I as soon as once more see plenty of deceptive claims each time I test my telephone.
Like I defined final month, these AI headlines are akin to a bookstore changing the covers of the books it places on show — solely right here, the “bookstore” is the information tab that seems while you swipe proper on the homescreen of a Samsung Galaxy or Google Pixel telephone, and the “cowl” is likely to be a AI-generated lie as an alternative of the reality.
For instance, Google’s AI claimed final week that “US reverses overseas drone ban,” citing and linking to this PCMag story for the information. That’s not simply false — PCMag took pains to elucidate that it’s false within the story that Google hyperlinks to!
From PCMag’s story, bolding theirs:
I noticed a headline saying that the drone ban was dropped. Is that true? No. Whereas it’s true the Commerce Division ended its efforts to limit DJI and different drones from import in Jan. 2026, it solely did so as a result of it could be redundant within the wake of the FCC actions. The Commerce Division is a separate entity from the FCC, and its proposed restrictions have been by no means put in place to start with. Some studies on the Commerce Division’s determination have deceptive headlines that might make you suppose the federal government did an about-face, so I don’t blame you for being confused.
What does the writer of that PCMag story suppose? “It makes me really feel icky,” Jim Fisher tells me over the telephone. “I’d encourage folks to click on on tales and browse them, and never belief what Google is spoon-feeding them.”
He says Google needs to be utilizing the headline that people wrote, and if Google wants a abstract, it could use those that publications already submit to assist search engines like google and yahoo parse our work.
Google claims it’s not rewriting headlines. It characterizes these new choices as “trending matters,” regardless that every “trending matter” presents itself as one in every of our tales, hyperlinks to our tales, and makes use of our photographs, all with out competent fact-checking to make sure the AI is getting them proper.
In some methods, Google’s present implementation isn’t fairly as dangerous because it was a month in the past. I’ve seen fewer examples of egregious clickbait, partly as a result of Google Uncover is now serving me fairly just a few unadulterated information tales alongside its AI ones — although it does minimize off their real headlines far too rapidly, making many robust to learn.
The AI can also be not restricted to roughly 4 phrases per headline, so I not see nonsense headlines like “Microsoft builders utilizing AI” or “AI tag debate heats.” (As an alternative, I often see tripe like “Fares: Want AAA & AA Video games” or “Dispatch offered thousands and thousands; few averted romance.”)
However Google’s AI has no clue what components of those tales are new, related, vital, or true, and it could simply confuse one story for an additional.
On December twenty sixth, Google advised me that “Steam Machine worth & HDMI particulars emerge.” They hadn’t. On January eleventh, Google proclaimed that “ASUS ROG Ally X arrives.” (It arrived in 2024; the brand new Xbox Ally arrived months ago.) On January twentieth, it wrote that “Glasses-free 3D tech wows,” introducing readers to “New 3D tech known as Immensity from Leia” — however linking to this TechRadar story about a wholly completely different firm known as Visible Semiconductor. I discovered one other that claimed to be a couple of GPU maker commenting on the RAM shortage; it linked as an alternative to a Digitimes story a couple of RAM maker.
I’m notably annoyed once I see bait-and-switch headlines on Verge tales, after all, and anxious they’re taking away our capability to market our personal work.
Google boiled down my colleague Jay Peters’ story about how RGB stripe OLED displays can unlock sharper text and more accurate colors to the boring “New OLED Gaming Screens Debut.” My story about letting you experience an immersive 3D demo of the Lego Smart Brick like you’re there with us at CES grew to become “Lego Good Play launches March 1,” a date that wasn’t information by the point Google wrote that! And Google AI determined to promote our big Verge Awards at CES 2026 story as “Robots & AI Take CES,” which was principally the other of our conclusion in that story.
And but, our new AI overlords aren’t successfully hunting down the worst human clickbait in trade for our fealty. One headline that Google’s AI didn’t overwrite was “Star Wars Outlaws Free Obtain Accessible For Much less Than 24 Hours” by Display screen Rant. Right here’s what the writer of that story reveals midway down the put up:
Simply the one code being given away when 1000’s of individuals are more likely to get entangled feels a bit stingy, though on the time of writing Ubisoft’s put up hasn’t precisely blown up, so that you is likely to be in with a shot.
Sure, Ubisoft gave away a single copy of a recreation on X, in a giveaway solely open to residents of the UK, but Google’s information bot determined that Display screen Rant’s FOMO clickbait was tremendous to serve up with out tweaks in any respect.
Right here’s the assertion from Google spokesperson Jennifer Kutz in regards to the Google characteristic:
We launched a brand new feature final yr in Uncover to assist folks discover matters which are coated by a number of creators and web sites. The characteristic features a useful AI-powered overview of the subject, a featured picture, and hyperlinks to associated tales. The overview headline displays data throughout a spread of web sites, and isn’t a rewrite of a person article headline. This characteristic performs nicely for person satisfaction, and we proceed to experiment with the UI to assist folks click on via and discover content material on the net.
Google declined our request for an interview to extra totally clarify the thought.
I don’t understand how broadly Google is displaying these “trending matters” AI headlines but, nevertheless it appears the corporate is testing them past the Google Uncover information feed, too. I’ve lately seen a few of them seem as push notifications to my telephone; tapping them takes me to a Google Gemini chatbot that makes an attempt to summarize a latest piece of reports.
Google adjustments like these are the most important cause The Verge now has a subscription, with out which we gained’t survive Google Zero.
Disclosure: Vox Media, The Verge’s guardian firm, has filed a lawsuit against Google, looking for damages from its unlawful advert tech monopoly.
