Overview
- Sam Altman announced on November 14 that ChatGPT now honors requests to avoid em dashes when the preference is saved in Custom Instructions.
- OpenAI clarified on Threads that the adjustment is a personalization option rather than a default removal of the punctuation.
- Users had long reported that the chatbot continued inserting em dashes even after explicit requests to avoid them.
- The em dash became a recognizable cue of AI-generated text, prompting some writers to change their punctuation to avoid suspicion.
- The tweak arrives as part of OpenAI’s broader push for user customization, including features like memory and custom GPTs, and Altman’s post drew significant online engagement.