According to Decrypt, new research from the Pew Research Center indicates that approximately one-third of websites launched following the release of ChatGPT were found to be primarily written by artificial intelligence. The study scanned nearly 500,000 webpages using a specialized AI detection tool.
The findings reveal a significant concentration of algorithmically generated content on .com domains, suggesting these sites may prioritize speed and volume over unique human authorship. Researchers noted that the prevalence of such material is increasing rapidly across the internet landscape, reflecting broader shifts in digital publishing strategies post-ChatGPT.
The Pew team utilized an AI detector to identify patterns characteristic of machine-generated text within this vast dataset. By focusing on nearly half a million pages, they were able to draw conclusions about how content creation has evolved since 2023 when ChatGPT became widely available. The study highlights that while human-written articles still exist in abundance, the sheer volume of AI-produced webpages is becoming difficult for casual observers to distinguish without technical tools.
This trend underscores a growing concern regarding transparency and authenticity on the open web. As more entities adopt automated writing solutions, particularly within commercial sectors represented by .com extensions, maintaining trust among readers becomes increasingly complex. The report does not specify exact percentages beyond the “one-third” estimate for AI-written content but emphasizes that this figure represents only those sites heavily reliant on machine assistance.
Implications suggest publishers may need to adjust their approaches toward disclosure or verification methods as automated writing gains traction globally. With nearly half a million pages examined, Pew provides valuable insight into how quickly digital ecosystems are transforming under the influence of large language models like ChatGPT and its successors.
