ChatGPT got rid of the habit that revealed AI in the text: OpenAI updated punctuation

OpenAI announced that its chatbot, ChatGPT, learned to no longer use the em-dash, which often reveals text written by artificial intelligence.

What is known

The "ChatGPT dash" appeared in school essays, letters, comments, social media posts, and even in advertising texts, causing suspicions that they were created by generative artificial intelligence.

Critics noted that excessive use of the dash distorted the style and "revealed" the chatbot. Users complained that even direct requests to ChatGPT not to use the dash did not work. The issue caused frustration and became a meme in communities tracking signs of AI text.

Now OpenAI assures that everything will change. In a post on X, the company's CEO Sam Altman called the update "a small but pleasant victory." If you specify in the chatbot not to use long dashes in personal instructions, it will begin to act according to your request.

In Threads, the company added: ChatGPT even apologized for "spoiling long dashes." Although by default the punctuation will not change, users will have more control over the response style.

Source: Sam Altman