Ever feel like your ChatGPT output has a… particular flavor? A certain punctuation mark that just keeps showing up, over and over again?
Well, I just saw this incredible post by Tianyu Xu that puts some hard numbers on this feeling, and honestly, it’s mind-blowing.
He decided to track his own ChatGPT usage for just one day. The results? It seems the model has a serious love affair with the em dash.
Here’s what Tianyu Xu discovered:
- 📊 The Count: A whopping 385 em dashes in a single 24-hour period.
- 😲 The Frequency: 36 out of his 55 chat sessions contained at least one em dash.
- ✍️ The Source: He noticed writing prompts produced the most, while image generation chats had the fewest.
So what’s the big deal?
As Tianyu brilliantly points out, the em dash is beautiful, but its overuse is quickly becoming a tell-tale sign of AI-generated content. It’s a tricky situation for writers who actually like using them!
Luckily, Tianyu shared three solid methods to help you avoid it:
1️⃣ Just Ask: Prompt ChatGPT not to use the em dash in its response.
2️⃣ Human First: Write the draft yourself and then use AI for revision and polishing.
3️⃣ Switch It Up: Try a different model, like Gemini, which may not have the same stylistic quirk.
I think he’s spot on. While it’s unfair to label any article with an em dash as AI-written, Tianyu’s findings are a great reminder to be mindful of these patterns to keep our writing fresh and human-sounding.
This is just a quick look at his awesome detective work. I highly recommend you read Tianyu Xu’s original post to see his full analysis and join the conversation!