The Pew Research Center analyzed nearly half a million English-language web pages for signs of text created by artificial intelligence. According to the summary of the study published by The Decoder, more than one-third of pages posted after the launch of ChatGPT contain such signs.
According to this data, commercial sites in the.com domain are ten times more likely to contain AI text than sites in the.edu or.gov domains. This indicates a difference between types of web platforms but does not explain the reasons for this gap.
The practical implication of the finding is the growing role of automated text in the public web environment. However, the available package is based on metadata and a synopsis from The Decoder, rather than the full text of the study or an independent verification of its results.
editorial commentary
Why it matters
If the findings are confirmed, newsrooms and researchers will need to more frequently verify the origin of web texts and distinguish automated content preparation. The next observable signal will be the emergence of comparable studies with open methodologies and independent verification. Significant uncertainty remains because the available package lacks the full text of the study and a description of the detection method.