ChatGPT won't add em dashes forcefully in your writing anymore
14 Nov 2025

OpenAI's CEO Sam Altman has announced a small but happy win in ChatGPT's behavior.

In a post on X, he said the AI model now obeys user instructions not to use em dashes.

The update has sparked an online conversation about punctuation habits, user frustrations, and writing preferences.

Many users have shared their experiences with the persistent em dash that ChatGPT used even when explicitly told not to.

Users share their experiences with ChatGPT's em dash usage
User reactions

The announcement by Altman has led to a flood of reactions from users who have had similar experiences.

One user said, "Told it to stop using em dashes... it replied with an em dash. Peak betrayal."

Another user expressed relief that they could now use their favorite punctuation mark without being accused of relying on unedited AI writing.

Altman's announcement sparks humor and relief among users
Humorous responses

The thread soon turned into a humorous exchange with users sharing their own experiences.

One user joked, "Huge update. Em dash removed. Now tell me: which punctuation is ruining your life today?"

Another user said, "Finally. The em dash was acting like the main character for no reason."

These comments highlight how users have often found the em dash overly prominent in AI-generated text.

Users express gratitude for ChatGPT's improved behavior
User appreciation

Some users expressed their relief at the update, with one saying, "Thanks for fixing this. Em dashes made me insane every time I asked it to edit my messages."

Another user jokingly asked, "But how will we know they are machines now?"

These reactions show that Altman's announcement has been well received by some in the community, and is seen by them as a step toward improving ChatGPT's writing capabilities.