OpenAI says it’s fixed ChatGPT’s em dash problem
OpenAI says ChatGPT will now remove the em dashes when you tell it to. The telltale sign that supposedly indicates text written by AI has banged upwards everywhere in recent months, including in school newspapers, emails, comments, customer service chats, LinkedIn posts, online forums, ad copy and more. The inclusion of the em dashboard has led to people criticizing these writers for being lazy and turning to an AI chatbot to do their work.
Of course, many have also advocated for the em dash, saying it has been a part of their writing long before LLMs adopted punctuation. However, the fact that chatbots could not avoid their use made the so-called “ChatGPT hyphen” a new objectionable addition to any text, even though they were not a reliable signal of content created by generative AI.
The problem had astonished OpenAI has been around for a while, as ChatGPT users were unable to get the chatbot to stop using the symbol even when they specifically asked not to.
Now OpenAI CEO Sam Altman says the company has resolved the problem. In a message on XAltman writes, “When 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 victory.”
The company explains in a post on Threads (where it forced ChatGPT to apologize for “ruining the em dashboard”) that ChatGPT will be better about not using the em dashboard if you instruct it not to do so via the custom instructions in your personalization settings. That means the em dash isn’t necessarily removed from the output by default, but at least you have more control over the frequency of its appearance.
