Google’s AI is rewriting headlines, and it’s already messy
Google is testing more AI inside its services, this time in the Discover feed.
According to The Verge, some users are now seeing AI-rewritten headlines instead of the original titles written by publishers.
In several cases, these new headlines were inaccurate or misleading.
One example involved an article from Ars Technica about a future Steam Machine device from Valve.
The real headline clearly said pricing had not been announced.
However, the version shown in Discover claimed the “Steam Machine price [was] revealed,” which was not true.
Google said this is only a small design test affecting a limited number of users.
The company explained that the aim is to make topics easier to scan before people click through to websites.
Even so, the test comes at a sensitive time for publishers, who have long raised concerns about how Google displays and repackages their content.
Here’s what you should know:
Google is testing AI-generated headlines and summaries inside Discover.
Some rewritten headlines have already contained factual errors.
AI Mode is being built directly into the main mobile search view.
Not the real title
The News Media Alliance has previously criticised Google’s AI tools for how they use news content.
At the same time, Google is pushing ahead with deeper AI integration in Search.
Its AI chatbot feature, known as AI Mode, is being tested directly alongside AI Overviews on mobile search.
Robby Stein confirmed the change on X, signalling that AI features are becoming more central to how Google Search works.
Google Discover said “plot twist” and meant it. - MG