A new Pew Research study has found that over a third of English-language webpages published since ChatGPT's November 2022 launch show signs of having been written or substantially edited by AI. Released on Thursday, the findings echo other research pointing to the rapid rise of AI-generated content online and follow a recent Cloudflare report showing bot web traffic has now overtaken human traffic, arriving sooner than expected. Together, the figures highlight how much of the web is increasingly produced by, and potentially even consumed by, automated systems rather than people.

To reach its conclusions, Pew drew on the Common Crawl web archive, gathering nearly half a million English-language webpages spanning roughly five years, and used AI-detection technology from Open Pangram to flag likely AI authorship. A random sample of 10,000 pages from July 2026 showed about 10% with significant signs of AI writing, but restricting the analysis to pages published after ChatGPT's release raised that figure to 35%. AI authorship rates varied sharply by domain type: .com sites showed roughly ten times the rate found on .edu or .gov domains (around 1% each), while .org domains sat at 4.6%. Pew also noted a rise in stylistic AI "tells", such as em dashes, Oxford commas and phrases like "it's not X, it's Y", though it cautioned that detection tools can misclassify genuinely human-written pages.