ChatGPT to remove em-dashes from AI generated output if asked

17 November 2025

OpenAI CEO Sam Altman, posting on X:

If you tell ChatGPT not to use em-dashes in your custom instructions, it finally does what it’s supposed to do!

It finally does what it’s supposed to do.

Am I to think AI agents were not supposed to include em-dashes in their input (and here I go) — unless said use was grammatically correct — all along? Were agents initially supplied a list of punctuation marks and told they could be used as they saw fit?

In other words, quite indiscriminately?

How unfortunate to think poor training of agents in the correct use of em-dashes, and Oxford commas, has lead to the perception that any text containing them is AI generated.

Here’s hoping other AI companies follow ChatGPT’s lead, and the rest of us can resume using em-dashes, and other maligned punctuation marks, as intended.