• 3 min read
Google’s AI Search Keeps People Off the Web
Google’s AI Overviews and AI Mode may be keeping users inside Google, while publishers face falling human traffic and rising bot activity.

Image: Gizmodo
Google’s push into AI Overviews and AI Mode is changing a basic bargain of the web: people used to search, click through to websites, and read pages supported by ads. Now, according to a New York Times report cited by Gizmodo, users increasingly get their answers from Google itself and never leave the results page.
A study from Growth Memo referenced by the Times found that 75% of people who use AI Mode stay there instead of switching back to the web. At the same time, a Cloudflare report from earlier this month said more than 50% of traffic on the Internet is now non-human. Put together, the trend is bleak for publishers: fewer humans arriving from search, while more impressions are served to bots.
That pressure comes as the web is already filling up with machine-made content. Gizmodo notes that a study from earlier this year found over 35% of new websites were AI generated. In that framing, the internet is becoming both AI-produced and AI-consumed.
Google disputes the darker interpretation. About a year ago, the company said search traffic remained strong and that claims to the contrary were “often based on flawed methodologies, isolated examples, or traffic changes that occurred prior to the roll out of AI features in Search.” A spokesperson also told the Times that new features in Google Search’s AI Mode surface relevant web links, an apparent attempt to push users back out to publishers.
Google also challenged the 75% figure directly, telling Gizmodo the study was not representative because participants were given “forced, unnatural tasks that likely influenced their behavior.”

Recommended reading
Starlink Mini 2 leak points to slower uplink speeds
Gizmodo’s own example shows why users may still drift deeper into Google’s chatbot interface. The publication tried asking whether any head of government before Andy Burnham had commonly gone by Andy. Google’s AI first produced quick answers naming Andry Rajoelina, Andrew “Andy” Holness, and Andy Clendennen.
Those answers were wrong, Gizmodo wrote: Andry Rajoelina is spelled with an r in his first name, Andrew Holness does not go by Andy, and Andy Clendennen appears to be fabricated. After the writer challenged the response in AI Mode, Google’s chatbot backed down.
Google told Gizmodo that its Search AI responses are “accurate the vast majority of the time,” and that mistakes are used to improve the system. But the larger shift may matter more than any single bad answer. As Gizmodo notes, citing remarks Google made in May at its developer conference, users now type queries that are three times longer than before, moving from keyword searches toward prompt-like requests.
That is a very different web behavior. Instead of asking whether someone has published a useful page, users increasingly ask Google to read the web on their behalf.
Culture Editor
Maya explores gaming, streaming, and the internet as a place where people actually live. From deep-dives into creator economies to the anthropology of digital communities, she tracks platform drama and cultural shifts so you don't have to. She believes the best tech stories are fundamentally about human behavior.
via Gizmodo


