Just weeks after Cloudflare reported that bot web traffic had surpassed human web traffic—a milestone reached earlier than the company estimated—a new Pew Research study released on Thursday reveals that more than one-third of web pages published after ChatGPT’s launch show signs of AI authorship. The report, which analyzes nearly half a million English-language web pages, provides a broad snapshot of how much of the web is now being written or substantially edited by artificial intelligence.

The Pew study focuses on what is being browsed, not on who or what is browsing the web, and much of the content is bots reading web pages written by other bots. To compile the report, Pew used the Common Crawl web archive to collect a vast set of pages from the past five or so years, starting a couple of years before ChatGPT’s November 2022 release. The team then applied Open Pangram’s technology to detect how many of the collected pages were likely written or heavily edited by AI.

Key Findings on AI Authorship

In a random sample of 10,000 web pages collected in July 2026, around 10% showed significant signs of AI authorship. However, Pew noted that a random sample would inevitably include older web pages published before AI writing tools existed, which couldn’t have been AI-written. To better assess AI-written content, Pew filtered out older web pages and focused only on those published after ChatGPT’s release. In that snapshot with older pages removed, signs of AI authorship were found in over one-third (35%) of the pages.

The report also found stark differences across domain types. URLs with a .com domain showed signs of AI authorship at around 10 times the rate of .edu or .gov domains, both of which were around 1% AI-authored. .org domains had only a 4.6% rate of AI authorship.

Caveats and Context

Pew acknowledged that Pangram, like other AI-detection tools, can misclassify pages as AI-written when they weren’t, making the analysis imperfect. Still, at scale, the data from the analysis is likely at least directionally correct. The study also found that other supposed tells of AI authorship had increased over the years, such as the use of em dashes, Oxford commas, and phrasing like “it’s not X, it’s Y.”

The findings corroborate other studies indicating that over a third of newer web pages are either written by AI or substantially edited by AI, painting a picture of a rapidly shifting online landscape where automated content is becoming the norm.