Pew Research analyzed Common Crawl data and found .com sites 10x more likely to have AI authorship than .edu or .gov.

Pew Research released a study Thursday finding that 35% of English-language web pages published after ChatGPT’s launch show signs of AI authorship. The analysis used Common Crawl to collect nearly half a million pages from the past five years and applied Open Pangram’s detection technology.

The report arrives after Cloudflare reported that bot web traffic overtook human web traffic, a milestone reached sooner than the company estimated. Pew’s data focuses not on who browses the web but on what is being browsed, and much of it is bots reading pages written by other bots.

For builders and operators, the findings signal that .com pages are 10 times more likely to show AI authorship than .edu or .gov pages. Since bot traffic now exceeds human traffic, builders should expect more AI-written content in the pages they fetch.

Pew cautions that AI detection tools can misclassify human writing, so the data is directionally correct but not exact. The study also found increased use of em dashes, Oxford commas, and phrasing like it’s not X, it’s Y as indicators of AI authorship, and future detection will require continuous adaptation.

What matters

• Pew Research found 35% of English-language web pages published after ChatGPT show signs of AI authorship.
• Builders should expect more AI-written pages in scraped data as bot traffic overtakes human traffic.
• Watch for AI detection limits; em dashes and Oxford commas now flag generated content.

Why it matters

Watch for AI detection limits; em dashes and Oxford commas now flag generated content.

This GenAI News article was prepared in original wording using reporting and materials published by TechCrunch AI. Source reference: https://techcrunch.com/2026/08/20/a-third-of-webpages-published-since-chatgpts-launch-show-signs-of-ai-authorship-study-finds/.

Drafted by the GenAI News review pipeline.