Overview
- OpenAI CEO Sam Altman said GPT-5.1 will honor a request in Custom Instructions to avoid em dashes, calling it a “small‑but‑happy win.”
- Users can add a clear rule in the “How would you like ChatGPT to respond?” field, such as “Do not use em dashes,” then save and start a new chat.
- Guides report improved compliance compared with earlier models, reflecting GPT-5.1’s better instruction-following on formatting choices.
- Outlets note the setting reduces but does not fully eliminate em dashes, with some users still seeing occasional slip‑throughs.
- Coverage highlights that the em dash remains a common tell in AI writing and that other models like Grok still rely on it heavily.