AI writing gives itself away in two ways: overusing em dashes where a period or comma belongs, and sounding too polished and thorough, like an academic email rather than how the actual person talks. The speaker even tells ChatGPT to skip em dashes to avoid detection.
One of the first dead giveaways that AI is being used is something called the em dash. An em dash is actually correct grammar, but it's that really long dash that very often in a sentence replaces dot, dot, dot, or in some cases, a comma. It's a long dash and it's technically called an em dash. AI loves to use the em dash. Most humans don't use that long em dash. So one of the things that I have done is I've told ChatGPT, hey, don't use the em dash when you're helping me write emails. Okay, that's kind of a dead giveaway. Alright, the second dead giveaway is when the answers you give, or the emails you send, are too well organized, they're too long, they give too much information, seems academic, seems like it was written by a professional. And if the person knows you, maybe it doesn't even sound like you. So you want to be pretty careful when you're using AI because it's going to be extremely thorough. It's going to contain absolutely no errors typically in terms of grammar and spelling. It's just too much. That's the second thing is when your emails are just too long and too thorough and don't sound like you. Okay.
