For example (and I am a native English speaker but I still make mistakes), https://chatgpt.com/share/d891f9ac-923b-47a8-8de9-ab73017ba9...
By starting from what you write (with your tone and words) it can more accurately maintain that tone and help you identify where you make mistakes and become more proficient in English.
Those are just for ChatGPT specifically, personally I know a ton of peculiarities from a wide variety of models that seem eerily consistent across prompts. You're very unlikely to accidentally use them though, much like spamming em dashes everywhere.
You can use whatever you like, but if you quack like a duck, people might think you're a duck.
So to me it seemed totally in character, so to speak.
Here's one to me classic example[1].
From https://en.wikipedia.org/wiki/Dash#Usage_in_AI-generated_tex... :
> In April 2025, Rolling Stone reported on the growing perception that the em dash is a hallmark of AI-generated writing, particularly by ChatGPT. The article noted how this idea spread through social media, where users began referring to it as the "ChatGPT hyphen" and how these users advised avoiding it to appear more human.