Users Report Difficulty Getting ChatGPT to Stop Using Em Dashes
Why does a single punctuation mark matter to so many people? Because the promise of a controllable AI often hinges on tiny details—like whether a sentence ends with a dash or a period. When OpenAI announced that ChatGPT could be instructed to drop em dashes, expectations rose.
Early adopters tried it in newsletters, academic drafts, and even casual blog posts, hoping for a clean, dash‑free output. But the reality proved less tidy. The model kept slipping in the long, horizontal strokes, despite prompts that explicitly banned them.
Some users resorted to repeated re‑phrasing, others to code‑level tweaks, yet the em dash persisted. The frustration grew from a curiosity about model behavior into a full‑blown annoyance, prompting public pleas and even cash offers for a solution. This tension between promised flexibility and actual stubbornness sets the stage for the following user testimony.
Users have since found it difficult to instruct ChatGPT to avoid using em dashes in a piece of text. "I pushed, pulled, rewired and begged, but nothing made it stop," one user wrote in a blog post. "$100 for anyone who can show me how to get ChatGPT to stop using em dashes.
It's driving me insane," said Chip Huyen, the author of the popular book 'AI Engineering', in a post on X a few months ago. Multiple factors have been attributed to ChatGPT's increasing use of em dashes. The model reflects patterns present in its training data, which includes a significant amount of modern prose that relies on the mark for pacing, emphasis and fluid transitions.
"If [AI] relied a lot on either magazine writing or blog writing, then those two styles were quite fond of the em dash," said Aileen Gallagher, a journalism professor at Syracuse University, in a statement to the Washington Post earlier this year. Em dashes also offer a compact way to connect or interrupt ideas without additional structure, making them efficient from a token-generation standpoint. The excess usage of em dashes in ChatGPT has also been a subject of jokes and memes on social media.
Some also say that this correlation is unfair. "It's annoying that em dashes have become the telltale sign someone used ChatGPT to generate the text," said one user on X. "I use them often in my emails and writing--probably incorrectly.
Now everyone assumes I'm putting everything through ChatGPT." Altman announced this update a day after OpenAI released the GPT 5.1 model.
Will the new setting finally silence the dash? OpenAI’s CEO Sam Altman announced on X that a user’s Custom Instructions can now suppress em dashes, calling the tweak a “small‑but‑happy win.” The option lives in the Personalisation tab of the settings menu, where users can specify stylistic preferences. For many, the change feels like a direct response to a flood of complaints, including a blog post in which a frustrated writer described endless attempts to coax the model away from the punctuation and even offered $100 for a solution.
Yet the announcement offers no data on how consistently the instruction is honoured across prompts or languages, and some users may encounter the dash despite the new toggle, especially if it conflicts with other system cues. Until broader feedback emerges, it remains unclear whether the fix fully resolves the annoyance or merely adds another layer to the customization menu. The update, however, does demonstrate OpenAI’s willingness to iterate on user‑driven quirks.
Further Reading
- Please make it stop using em dashes! - OpenAI Developer Community
- ChatGPT — get away from my em dash - Winnipeg Free Press
- AI embraces the em dash — Its users? Not so much! - Evidence Explained
- The Em Dash Responds to the AI Allegations - McSweeney's
More in LLMs & Generative AI
Anthropic explains reinforcement learning metric for Claude’s wokeness
Anthropic has finally opened the black box around what it calls Claude’s “wokeness” score.
Hyperlink Agent Search on NVIDIA RTX PCs doubles LLM inference speed
Here's the thing: Hyperlink Agent Search just landed on NVIDIA RTX‑powered PCs, and the rollout...
LinkedIn scales AI people search to 1.3 B users, lifts non‑degree hires 10%
LinkedIn has taken its AI‑driven recruiting tools from a niche experiment to a platform that now...
PhonePe partners with OpenAI to bring ChatGPT to millions of Indian users
PhonePe is rolling out a new feature that will embed OpenAI’s ChatGPT directly into its payment...
User Feedback Drives Evaluation of Indian AI Models in New Indic LLM‑Arena
The launch of Indic LLM‑Arena marks a rare attempt to gauge how well large language models...
Common Questions Answered
How can users suppress em dashes in ChatGPT according to the article?
OpenAI added a setting in the Personalisation tab of the Custom Instructions menu that lets users specify stylistic preferences, including disabling em dashes. This option was announced by Sam Altman as a “small‑but‑happy win” and is intended to address the flood of user complaints.
What challenges did early adopters face when trying to stop ChatGPT from using em dashes?
Early adopters reported that despite explicitly instructing the model, it kept inserting em dashes in newsletters, academic drafts, and casual blog posts. Users described repeatedly pushing, pulling, rewiring, and begging the model, with one user even offering $100 for a solution.
Who publicly expressed frustration about ChatGPT’s use of em dashes, and what did they say?
Chip Huyen, author of the popular book 'AI Engineering', posted on X that the em dashes were driving him insane and offered $100 to anyone who could stop the model from using them. His comment highlighted the broader dissatisfaction among writers over the model’s punctuation habits.
What does the article suggest about the significance of a single punctuation mark like an em dash for AI controllability?
The article argues that tiny details such as whether a sentence ends with an em dash or a period can affect the perceived controllability of AI outputs. Users view the ability to manage such punctuation as a measure of how finely they can steer the model’s style.