Back to feed
News Story
THE DECODER
1 sources

Pew Study Confirms Sharp Rise of AI-Written Text on the Web Since ChatGPT's Launch

The Pew Research Center analyzed nearly half a million English-language web pages and found that more than a third of pages published since ChatGPT's launch show signs of machine-written text. Commercial .com sites are ten times more likely to contain AI content than .edu or .gov domains.

SynthePulse Insight · AI deep readingMembers

AI Text Now One-Third of New Web Pages: Pew Research Reveals the Quiet Transformation of Online Writing

Version 1 · 1 source

A Pew Research Center analysis of nearly 500,000 English-language web pages shows that since ChatGPT's release, more than one-third of new pages show signs of AI generation, with commercial sites ten times more likely than educational or government sites. The findings both confirm the explosion of AI writing and expose the limitations of detection tools and the ambiguity of definitions.

  • Pew analyzed nearly 500,000 English-language web pages and found that since ChatGPT's release, more than one-third of new pages show signs of AI generation.
  • In the July 2026 sample, about 10% of pages showed clear AI authorship traces; .com domains had about one in ten, while .edu and .gov had only about 1%.
  • AI-favored words like 'delve' and 'pivotal' doubled in frequency, em-dash usage doubled, and Oxford comma usage increased by 63%.
Open section navigationCore Findings: Significant Growth in AI Text

Core Findings: Significant Growth in AI Text

The Pew Research Center analyzed nearly 500,000 English-language web pages from the Common Crawl web archive, using the AI detection tool Open Pangram. In the July 2026 sample, about 10% of pages showed clear signs of AI authorship.

When counting only pages published after ChatGPT's release (late 2022), that share jumps to more than one-third. The study notes that the share of AI-generated web content has steadily climbed since ChatGPT's debut.

Free for now

Read the full analysis

4 more sections of analysis, plus the full takeaway

Loading

Credibility boundary

This report is based on THE DECODER's summary of the Pew study. Core data (such as sample size and percentages) come from the Pew study, but the original report was not directly cited, so it is marked as source_claim. The assessment of detection tool capabilities and the definitional controversy include the author's personal experience and are inferential.

Primary report

THE DECODER

Primary source