# ChatGPT's Em Dash Habit: A Training Artifact or Design Choice?

**URL:** <https://community.openai.com/t/chatgpts-em-dash-habit-a-training-artifact-or-design-choice/1115873>\
**Category:** Community\
**Tags:** chatgpt\
**Created:** [February 8, 2025, 2:45pm UTC](https://community.openai.com/t/chatgpts-em-dash-habit-a-training-artifact-or-design-choice/1115873 "2025-02-08T14:45:38Z")\
**Posts on this page:** 1\
**Showing post:** 5

<div class="post-metadata">

**Author:** ![smm.tim](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/smm.tim/32/626081_2.png) [@smm.tim](https://community.openai.com/u/smm.tim)\
**Post date:** [May 15, 2025, 11:24am UTC](https://community.openai.com/t/chatgpts-em-dash-habit-a-training-artifact-or-design-choice/1115873/5 "2025-05-15T11:24:04Z")

</div>

I’ve noticed that even when I specifically ask ChatGPT not to use em dashes, it still includes them in its responses.

**Could any of the developers shed some light on why it doesn’t always follow writing guidelines or user prompts?**

I’m curious to understand how this behavior works and if there’s a way to improve consistency.

---

_[View the full topic](https://community.openai.com/t/chatgpts-em-dash-habit-a-training-artifact-or-design-choice/1115873)._