# 10% of the web now shows signs of AI authorship Pew study finds AI traces in one-third of pages published post-ChatGPT - **Source:** [The Next Web](https://brifing.si/source/afe24b20-b275-4c19-b9a6-8bb3c8babae7) - **Published:** 2026-08-20 - **Original article:** - **On Brifing:** https://brifing.si/s/97a2d7a2-b7f4-4714-b33f-86b372355320 ## Prevalence of AI authorship Pew Research's analysis of 10,000 randomly selected web pages in July 2026 revealed that about 10% showed strong signs of AI authorship. For pages published after ChatGPT's 2022 launch, this figure jumps to one-third. Commercial domains (.com) carry the majority of AI-generated content, reflecting cost-driven content production. ## Distribution by domain type The study found significant variations in AI authorship across domain types. While 10% of .com pages showed AI markers, the rate was 4.6% for .org and approximately 1% each for .edu and .gov. This distribution paints a clear picture of where writing serves as a cost versus a core purpose in different sectors. ## Detection methodology Pew analyzed 500,000 English pages from Common Crawl archives (January 2021-July 2026) using Open Pangram, the detector Substack employs to catch machine-written newsletters. While individual documents may be misclassified, the system provides statistically reliable patterns across hundreds of thousands of pages. ## Linguistic fingerprints of AI writing The research identified distinct linguistic patterns in suspected AI texts. Em dash usage has doubled since 2023, Oxford commas increased by 63%, and words like 'delve' and 'pivotal' frequently appear alongside the 'not just X, but Y' construction - all becoming telltale signs of machine-generated content. ## Platform countermeasures LinkedIn claims its system identifies generic content with 94% accuracy, while arXiv now bans researchers submitting unchecked AI text. These measures reflect platforms' efforts to fortify their infrastructure against the AI content wave while maintaining content quality standards. ## Visibility versus volume Separate Graphite research shows AI-generated articles approaching 50% of new publications, yet human-written pieces still dominate Google results and academic citations. This suggests AI content may be growing in volume but remains limited in actual impact and authority. ## Industry transformation AI authorship has become a default reality - any page published in 2026 likely had some model involvement. The industry response focuses on providing users with verification tools, creating a new market for content authenticity solutions in the AI era. ## Content production paradigm shift Pew's findings confirm AI's permanent role in web content creation, particularly revolutionizing cost-driven commercial writing. However, low adoption in education (.edu) and government (.gov) sectors suggests human authorship retains its value where accuracy and accountability matter most. ## Sources covering this story - https://thenextweb.com/news/pew-ai-authorship-web-pages-study — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._