As AI infuses its way into the mainstream, people who write a lot (emails, pitches, papers) are trying out chatbots like ChatGPT to assist.
But although the technology has advanced, it still has a long way to go — and sometimes, certain words and sentence structures just don’t feel quite right.
And according to an AI expert, there’s one word that’s a dead giveaway your email was written by ChatGPT.
Related: ChatGPT: What Is It and How Does It Work?
Paul Graham, the co-founder of startup accelerator company Y Combinator, posted on X this week that using the word “delve” is a red flag that the work was most likely written by ChatGPT.
My point here is not that I dislike “delve,” though I do, but that it’s a sign that text was written by ChatGPT. pic.twitter.com/2i3GUgynuL
— Paul Graham (@paulg) April 7, 2024
He referenced a chart by researcher Philip Shapira, which shows the usage of the word “delve” from 1990 to 2024 in published papers and articles jumps to nearly 18,000 instances in the last four years.
Another data tool, AI Phrase Finder, put “delve” as the No. 9 (out of 10) most common word used by ChatGPT.
“If you make demands for a piece of text to be comprehensive in coverage, then you can’t exactly blame ChatGPT for using a word like ‘delve,’ which carries connotations of pulling your sleeves up and digging into the depths of a subject matter,” AI Phrase Finder said. “Still, it’s undoubtedly a word that’s overused by AI.”
Related: The 7-Step ChatGPT Formula for Peak Productivity and Profit
According to Merriam-Webster, delve is a verb that means “to make a careful or detailed search for information” and “to dig or labor with.”
