According to OpenAI, ChatGPT will now stop using em dashes if you specifically request it. This punctuation mark, which many believe is a clear indicator of AI-generated writing, has become increasingly common in recent months, showing up in everything from student essays and emails to online comments, customer support chats, LinkedIn updates, forum posts, advertisements, and more. The frequent use of em dashes has led some to accuse writers of relying on AI tools instead of crafting their own work.
On the other hand, plenty of people have defended the em dash, claiming it’s been a staple in their writing long before large language models began using it. Still, the fact that chatbots seemed unable to avoid this punctuation turned the so-called “ChatGPT hyphen” into a new point of criticism, even if it wasn’t a foolproof way to spot AI-generated content.
This issue puzzled OpenAI for quite a while, as users found that ChatGPT would continue to insert em dashes even when specifically instructed not to do so.
Recently, OpenAI CEO Sam Altman announced that the issue has been resolved. In a post on X, Altman shared, “If you tell ChatGPT not to use em-dashes in your custom instructions, it finally does what it’s supposed to do,” describing the fix as a “small-but-happy win.”
In a Threads post (where ChatGPT was made to apologize for “ruining the em dash”), the company clarified that ChatGPT will now better avoid using em dashes if you specify this preference in your personalization settings. While this doesn’t mean the em dash will be removed by default, it does give users more influence over how often it appears in generated text.
