Artificial intelligence is transforming the internet in two ways: attracting billions of visits to AI platforms while automated systems increasingly account for the traffic moving between websites.
New research suggests the consequences could extend far beyond search engines, changing how online businesses attract audiences and generate revenue.
An analysis by OneLittleWeb found that four major AI platforms – ChatGPT, Gemini, Claude and Grok – gained a combined 35.6 billion annual web visits when comparing July 2024 to June 2025 with July 2025 to June 2026.
The study examined estimated traffic across 10,171 websites, using data from Ahrefs and Semrush.
ChatGPT recorded the second-largest absolute increase of any website analyzed, gaining 22.2 billion visits, behind YouTube’s 25.6 billion.
How AI is changing web traffic
| Platform | Previous visits | Latest visits | Change | Change (%) |
|---|---|---|---|---|
| YouTube | 570.1bn | 595.7bn | +25.6bn | +4.5% |
| ChatGPT | 43.6bn | 65.8bn | +22.2bn | +51.0% |
| Gemini | 1.5bn | 8.7bn | +7.3bn | +494.1% |
| Claude | 1.1bn | 5.1bn | +4.0bn | +365.7% |
| Grok | 0.53bn | 2.7bn | +2.1bn | +398.4% |
| Yahoo Japan | 25.9bn | 22.9bn | -3.0bn | -11.5% |
| Baidu | 8.5bn | 5.4bn | -3.1bn | -36.5% |
| Google.co.uk | 5.4bn | 1.8bn | -3.6bn | -66.1% |
| Bing | 48.6bn | 24.7bn | -23.9bn | -49.1% |
(Source: OneLittleWeb, using Ahrefs and Semrush estimates. Figures compare July 2024-June 2025 with July 2025-June 2026. Visits are estimated, not unique users. Some figures are rounded.)
Meanwhile, several established search engines experienced substantial declines.
AI platforms rise as search engines lose traffic
According to OneLittleWeb, Bing recorded the largest decline in the dataset, losing 23.9 billion visits, almost matching ChatGPT’s increase.
Google.co.uk lost 3.6 billion visits, Baidu lost 3.1 billion and Yahoo Japan lost 3 billion.
Together, those four search domains lost approximately 33.5 billion visits.
The figures do not establish that users moved directly from search engines to AI platforms. They also measure individual websites rather than all the services offered by their parent companies.
Nevertheless, they illustrate how rapidly the distribution of online attention is changing.
Other established destinations experienced declines, including Wikipedia, which lost 9.8 billion visits, and Quora, which lost 3.6 billion.
The Dead Internet Theory meets measurable reality
These developments invite comparisons with the “Dead Internet Theory”, an internet conspiracy theory suggesting that much of the web’s activity and content is generated by automated systems rather than humans.
Its more extreme claims about coordinated manipulation remain unsubstantiated. However, the growing dominance of automated traffic is measurable.
According to Thales’ 2026 Bad Bot Report, bots accounted for approximately 53 percent of web traffic in 2025, compared with 47 percent attributed to humans.
Not all bots are malicious, and not all are powered by large language models. They include search crawlers, monitoring systems, malicious scrapers and increasingly sophisticated AI agents.
Separate research by Cloudflare found that automated crawling initiated by users’ interactions with AI services increased more than fifteenfold during 2025.
The distinction matters: AI is not simply attracting people to new websites. It is also changing which machines visit existing websites, and why.
A changing business model for the web
For publishers and online businesses, the implications could be considerable.
Traditional search engines have historically directed visitors to websites where publishers could generate advertising revenue, subscriptions or commercial inquiries.
AI services increasingly provide answers within their own interfaces, potentially reducing the need to visit the original source.
At the same time, automated crawlers can consume website content without generating the human engagement on which many online businesses depend.
OneLittleWeb’s findings do not directly measure this economic impact, but they illustrate the changing balance between established information destinations and AI platforms.
From human audiences to machine customers
The next stage could involve AI agents doing more than retrieving information.
Automated systems are beginning to compare products, navigate websites and complete transactions on behalf of users.
That creates opportunities for businesses whose services can be discovered and accessed by machines, but raises questions about attribution, advertising and the value of human attention.
The internet is unlikely to become literally “dead”. People will still create, communicate, shop and seek information.
But the relationship between human users, websites and automated systems is changing.
For businesses that have spent decades optimizing websites to attract human visitors, the challenge may increasingly be to serve two audiences: people who want information and machines that retrieve it on their behalf.
Main image courtesy of Easy-Peasy.ai
