Over a third of webpages created after ChatGPT’s debut show signs of AI writing, study finds

A Pew Research Center study of over half a million English-language webpages found that 35% of pages created after ChatGPT's 2022 launch show signs of AI writing, compared with 10% of a random sample of 10,000 pages collected in July 2026. The share of AI-written text is lower on .edu and .gov domains and highest on .com pages, and the trend accelerated with the arrival of other chatbots like Claude and Gemini.

Over a third of webpages created after ChatGPT’s debut show signs of AI writing, study finds See more of our coverage in your search results. Add The New York Post on Google https://www.google.com/preferences/source?q=nypost.com More than a third of all webpages created after the launch of ChatGPT show signs of AI writing https://nypost.com/2026/08/19/world-news/students-who-use-ai-score-great-on-homework-but-bad-on-tests-study/ , according to a new study. The Pew Research Center investigated over half a million English-language webpages from the last five years — finding a sharp rise in machine-written text following the launch of ChatGPT in 2022. Researchers used an AI detection tool called Open Pangram to analyze a random sample of 10,000 webpages collected in July 2026 and found that 10% showed signs of AI authorship, according to the report published last week. https://www.pewresearch.org/data-labs/2026/08/20/how-much-of-the-internet-is-written-with-ai/ But the share climbed sharply when researchers analyzed only webpages created after ChatGPT’s debut. More than one-third — 35% — of the post-ChatGPT pages showed signs of AI authorship, echoing other studies suggesting that artificial intelligence is shaping a large part of the modern web. The trend began after ChatGPT’s launch and accelerated further with the arrival of additional popular chatbots, including Claude and Gemini. AI-written text still remains far less common on webpages with “.edu” or “.gov” domains and is most commonly found on “.com” pages, the study found. Though AI detection models are not foolproof, they can expose certain patterns associated with machine-generated writing, including specific punctuation, words, and phrases, researchers said. The findings come as additional studies revealed that AI chatbots have become increasingly commonplace in Americans’ lives. A Pew Research Center study released in June https://www.pewresearch.org/internet/2026/06/17/americans-and-ai-2026-chatbots-smart-devices-and-views-on-impact/ revealed that over half of US adults now report using chatbots. Roughly 38% of employed adults also reported using chatbots for work. Statistics also show that young adults under age 30 are more wary of AI’s potential negative impacts on society than older adults.