OpenAI has officially resolved a persistent issue where ChatGPT consistently ignored user instructions to avoid using em dashes. The update, announced by CEO Sam Altman, grants users the ability to finally suppress the punctuation mark that has become a hallmark of AI-generated content across emails, academic papers, and digital marketing copy.
The End of the “ChatGPT Hyphen”
For months, the unfiltered use of em dashes became a primary indicator of text generated by large language models. This stylistic tic appeared ubiquitously, leading to widespread criticism and accusations of laziness among users who relied on the chatbot for professional and personal communications. Despite its historical legitimacy in formal writing, the chatbot’s inability to suppress the symbol turned it into a target of intense public scrutiny.
Technical Hurdles and User Frustration
The technical struggle to curb this behavior had persisted for some time. Users frequently reported that ChatGPT would disregard specific prompts requesting the exclusion of em dashes, leaving them stumped by the model’s stubborn adherence to its default punctuation style. This limitation undermined the platform’s utility for those seeking to produce content that does not immediately signal AI involvement.
New Customization Controls
Addressing the feedback, CEO Sam Altman confirmed the fix via a post on X, describing the update as a “small-but-happy win” for platform usability.
Small-but-happy win:
If you tell ChatGPT not to use em-dashes in your custom instructions, it finally does what it's supposed to do!
— Sam Altman (@sama) November 14, 2025
According to the company’s official statement, the fix is integrated into the “custom instructions” feature within personalization settings. While this change does not strip the em dash from ChatGPT’s default output, it provides users with the necessary control to dictate the frequency and presence of the punctuation in their generated content, ensuring the model adheres to specific style preferences.