A study from Pew Research released on Thursday states that 35% of web pages published since ChatGPT‘s launch show signs of AI authorship.
The report confirms earlier findings that much of the internet’s newer content is either written by artificial intelligence or substantially edited by it. This data arrives shortly after internet infrastructure firm Cloudflare announced that bot traffic had surpassed human traffic, a milestone reached sooner than the company anticipated.
Pew’s research focuses on what is being read rather than who is reading it. The analysis suggests a significant portion of the web consists of bots accessing pages generated by other bots.
To compile the figures, the firm used the Common Crawl archive to gather nearly 500,000 English-language pages from the last five years. This period began a couple of years before ChatGPT launched in November 2022. Open Pangram technology identified which pages were likely written or heavily edited by AI.
In a random sample of 10,000 pages collected in July 2026, roughly 10% displayed significant signs of AI authorship. However, this sample included older pages published before AI writing tools existed, meaning those could not have been AI written.
Pew removed these older entries to focus strictly on content published after ChatGPT’s release. In this filtered snapshot, over one-third of the pages showed AI authorship indicators.
Domain types showed distinct differences. .com URLs displayed AI authorship signs at roughly 10 times the rate of .edu or .gov domains, both of which hovered around 1%. .org domains recorded a 4.6% rate of AI authorship.
This analysis is not perfect. Pangram, like other detection tools, can misclassify human-written pages as AI generated. Yet, at this scale, the data likely points in the right direction.
The study also noted that other markers of AI writing have grown over time. These include increased use of em dashes, Oxford commas, and phrasing such as “it’s not X, it’s Y”.
What it means
For writers and site owners, the shift is clear. Content published after late 2022 is now dominated by machine generation on commercial domains. The rise in bot traffic suggests that the web is increasingly a loop of automated systems consuming automated content, leaving little room for human curation or original reporting.
