A study of online content finds that the quantity of AI-generated articles now exceeds the number posted by human authors. But it's been a close-run thing for about 12 months, according to their sample of 65,000 English-language online items, suggesting the impact of ChatGPT and other Large Language Models has plateaued. Moreover, they suspect that the AI-generated article are largely unread by humans.

We hypothesize that this is because practitioners found that AI-generated articles do not perform well in search, as shown in a separate study.

They admit that that accurate detection of AI is difficult, and in particular that "a human in the loop" makes it moreso: "our study did not evaluate the prevalence of content created using this strategy." So you might say the study was of material

See Full Page