Over the past few months, this punctuation, once seen as a hallmark of AI-generated text, has appeared widely in school papers, emails, comments, customer service chats, LinkedIn posts, online forums, ad copy, and more.

Its frequent use led some readers to accuse writers of relying on AI rather than crafting the content themselves.

Many writers, however, have defended the em dash, noting it has long been a natural part of human writing.

Still, the persistence of em dashes in AI-generated text had earned it the nickname “ChatGPT hyphen,” making it a controversial marker, even though it was never a reliable indicator of AI authorship.

The issue had long frustrated ChatGPT users, who couldn’t get the AI to stop using em dashes, even when explicitly instructed.

Now, OpenAI CEO Sam Altman says the problem is resolved. In a post on X, he wrote, “If you tell ChatGPT not to use em-dashes in your custom instructions, it finally does what it’s supposed to do,” calling the update a “small-but-happy win.”

In a Threads post, where ChatGPT jokingly “apologized for ruining the em dash”, OpenAI said users can now limit em dash use by setting custom instructions in their personalization settings.

It won’t remove them entirely by default, but gives users more control over their frequency.