Sam Altman announces ChatGPT no longer forces em dashes into your writing: Here’s why users hated it

Sam Altman has announced what he calls a “small but happy win”. His reason? If you tell ChatGPT not to use em dashes, it finally listens. This might sound like a tiny victory in the world of artificial intelligence, but anyone who has spent time with ChatGPT knows this is actually a moment worth celebrating.

For years, users have begged the model to stop sprinkling em dashes everywhere. You could ask it to write a note, a poem, a news story, or even a message for your boss. No matter what you requested, the model often managed to slip in that long dramatic line in the middle of a sentence. It was the punctuation version of photobombing. Once you saw it, you knew exactly which machine had been at work.

The funny part is that humans have developed a very sharp em dash detector. People can now read a paragraph and instantly say, “AI wrote this.” Why? Because humans do not use em dashes with this kind of energy. They use them sparingly, almost shyly. ChatGPT used to treat them like a personal signature.

But here is the interesting twist. Before the ChatGPT era, em dashes had a completely different personality. Writers used them to add dramatic pause or to interrupt a thought. Novelists used them to show a character trailing off. Journalists used them when a full stop felt too harsh and a comma felt too soft. They were a tool, not a spotlight. Most readers barely noticed them.