OpenAI says ChatGPT will now avoid using em dashes when users request it, addressing a long-standing complaint that the punctuation mark had become a telltale sign of AI-generated writing.
The em dash — jokingly nicknamed the “ChatGPT hyphen” — has become widely associated with generative AI output across school assignments, emails, customer service chats, and online posts. Many users previously reported that the model continued to insert em dashes even when explicitly instructed not to.
OpenAI CEO Sam Altman announced the fix on X, calling it a “small-but-happy win.” The company explained that users can now disable em dashes through ChatGPT’s custom instructions, giving them more control over punctuation style. The model will still use em dashes by default unless instructed otherwise.
