<meta charset="UTF-8"> <meta name="viewport" content="width=device-width, initial-scale=1.0"> <title>HF Demo</title> <style> body { margin: 0; font-family: Arial, sans-serif; } header { background-color: #2c3e50; color: white; text-align: center; padding: 20px; font-size: 28px; font-weight: bold; } </style>

Pew says 10% of the web now shows signs of AI authorship

Pew: 10% of Web Content Now Shows Signs of AI Authorship

Pew classified nearly 500,000 pages from Common Crawl and found the AI share concentrated on commercial domains, at 10% of .com against 1% of .gov. (August 20, 2026)

About a tenth of the web now displays strong indications of having been written by a machine, according to Pew's July 2026 analysis of a random sample of 10,000 pages.

A more striking figure is that over a third of pages published since ChatGPT’s release in November 2022 exhibit these AI authorship signals.

Methodology:

Pew used Open Pangram, a Substack-adopted detector, to classify 500,000 English-language pages from Common Crawl between January 2021 and July 2026.

Key Findings:

  • Domain Breakdown: 10% of .com pages, 4.6% of .org, roughly 1% each of .edu and .gov showed AI authorship.
  • Writing Trends: Em dash use has doubled since 2023, Oxford commas are up 63%, and words like "delve" and "pivotal" are frequently used alongside phrases like "it’s not just X, it’s Y."

Industry Response:

Platforms are already integrating AI detection into their systems. LinkedIn claims 94% accuracy in identifying generic content, and arXiv now bans researchers who submit AI-generated text without human review.

While AI-generated volume is growing, human-written pieces still dominate Google results and AI citations. However, the default assumption for newly published content is now that it has been touched by a language model. Tools for detecting AI-generated content are increasingly available.

Leave a Reply