OpenAI finally fixes ChatGPT's 'em dash' problem

OpenAI has finally fixed ChatGPT’s em dash problem, according to the company’s CEO, Sam Altman. Over the weekend, Altman revealed the update in a post on X, calling it a “small-but-happy win.”

He wrote:

If you tell ChatGPT not to use em-dashes in your custom instructions, it finally does what it’s supposed to do.

For those unaware, the punctuation mark has become infamous as a sign of AI-generated text. Its frequent use led some to criticize writers for relying on AI rather than their own words.

However, there’s a section that defends the use of the em dash as a normal part of writing. Until now, users couldn’t get ChatGPT to stop using it, even when explicitly asked.

According to OpenAI, the change works through the custom instructions in the personalization settings. Users can now tell ChatGPT to avoid the em dash, giving them more control over its appearance.

Do note that the update doesn’t remove em dashes by default, but it ensures the AI respects user preferences when instructed.

As far as the company news is concerned, OpenAI also introduced GPT-5.1, the new and upgraded model, comprising GPT-5.1 Instant and GPT-5.1 Thinking. The company describes the new model as an “upgrade” to GPT-5 that makes ChatGPT smarter, faster, and more fun to talk to.