I think the author asked ChatGPT to write an article for him. The study he links to does not confirm any of the statistical claims he makes even though the article says it does.