Summary
- Pew Research examined about 490,000 English-language web pages using the Common Crawl database; 10% of a July 2026 sample indicated substantial AI authorship.
- When focusing solely on content published after the launch of ChatGPT in November 2022, this percentage rises to 35%.
- Websites with a '.com' domain show signs of AI authorship at a rate approximately ten times higher than '.edu' or '.gov' domains, and about twice that of '.org' domains.
The prevalence of AI-generated content on the internet is striking. A recent study from the Pew Research Center reveals that 10% of English-language webpages exhibit notable signs of AI authorship. This figure escalates to 35% when only considering pages published since the introduction of ChatGPT in November 2022.
Myriad: When will OpenAI release GPT-6? Make your prediction here.This analysis was conducted by examining approximately 490,000 pages from the Common Crawl web archive, covering the period from January 2021 to July 2026, and applying the Open Pangram AI detection model to the text.
AI content is not uniformly distributed across the web. Pages on '.com' domains show AI authorship indicators at a rate about ten times that of '.edu' and '.gov' domains, which hover around 1% each, and approximately double that of '.org' sites, which have a 4.6% rate. In 2021, the rates across all domain types were nearly the same.
Pew's methodology involved detecting AI-generated text using statistical patterns rather than specific words or phrases. However, certain indicators have surged since 2023: the use of em dashes has doubled, the frequency of Oxford commas has increased by 63%, and AI-preferred terms like "delve," "interplay," and "testament" have more than doubled in usage. Notably, the construction known as "negative parallelism"—phrases such as "it's not just X, it's Y"—has nearly tripled since 2023, although it remains uncommon overall. Decrypt has previously identified these indicators, and Merriam-Webster declared "slop" as its word of the year in December, highlighting the impact of AI on language.