OpenAI fixes ChatGPT's em dash problem; 'small-but-happy-win' says Sam Altman
OpenAI’s latest tweak to ChatGPT is like finally teaching your robot butler not to overuse exclamation points — it’s a small fix, but one that makes a world of difference in how natural the output feels. Sam Altman’s ‘small-but-happy win’ quip is spot-on: after months of users begging the AI to stop peppering every sentence with em dashes, the chatbot now actually listens. It’s a victory for writers everywhere who want their AI-assisted prose to pass as human, not as a textbook example of ChatGPT’s signature style.
This update is part of OpenAI’s broader push to make ChatGPT more customizable and user-friendly. By letting users control punctuation habits like em dash usage through custom instructions, OpenAI is addressing a major pain point for professionals, academics, and creatives who rely on the tool. While other AI tells — like clichéd phrases — remain, this fix is a step toward more authentic, personalized AI-generated content. The change also reflects growing awareness of how AI writing quirks can affect credibility and perception in professional and academic settings.
