AI News Feed
Market watch
Large Language Models

Pew Research Finds AI Traces in One-Third of Post-ChatGPT Web Pages

A Pew Research analysis of 500,000 pages found AI authorship signs in 10% overall and over one-third of pages published after ChatGPT's release.

The analysis, conducted by a senior data scientist at Pew Research, used an AI detection tool to examine 500,000 English-language webpages from the past five years. Across the entire sample, 10% of pages showed significant signs of AI authorship. The share rose sharply when only newer content was considered: in a July 2026 snapshot, more than one-third of pages published after ChatGPT's launch displayed such signs.

The distribution of AI-generated text is uneven across the web. At the time ChatGPT was first released, linguistic patterns associated with AI appeared at similar rates across major top-level domains, including .com, .org, .edu and .gov. By 2026, however, roughly one in ten .com pages showed signs of AI authorship — about double the share on .org domains, which stood at 4.6%, and ten times the rate on .edu or .gov domains, each around 1%.

The researchers also identified measurable changes in online writing style by comparing current internet text with a 2023 baseline. Em dashes now appear twice as often, while Oxford comma usage has increased by 63%. Words more commonly favored by AI, such as 'delve,' 'interplay' and 'testament,' have more than doubled in frequency. The use of negative parallelism — constructions like 'it's not just X, it's Y' — has nearly tripled.