AI News Feed
Market watch
World

Over a Third of Web Pages Published Since ChatGPT Launch Show AI Authorship, Pew Study Finds

A Pew Research Center study finds that over 35% of web pages published since ChatGPT's launch show signs of AI writing, with commercial sites leading the trend.

The study, which used automated tools to analyze a sample of 10,000 web pages drawn from the Common Crawl project, found that 35% of pages published since ChatGPT's debut in late 2022 showed signs of AI authorship. Across the entire sample, which dates back to the early days of the World Wide Web, about 9.6% of pages appeared to have been written by AI.

According to the report, the majority of suspected AI-generated pages published this year are on .com domains, typically commercial sites. In contrast, only 1% are on .edu domains (US educational institutions) and 0.8% on .gov domains.

The researchers also identified common features of AI-generated text. The em dash appears twice as frequently in pages published since 2023, while use of the Oxford comma has risen by 63%. Certain vocabulary, including "delve," "interplay," and "testament," has more than doubled in usage, and a negative parallel structure ("it's not just X, it's Y") has nearly tripled. The report authors caution that these are not solitary indicators, since humans also use such constructions.

The data suggests that pages from .edu or .gov domains are overall more likely to have been written by a human than by AI.