ChatGPT Em Dash Fix: OpenAI Resolves Issue

ChatGPT Now Avoids Em Dashes on Request

OpenAI has announced that ChatGPT can now be instructed to refrain from using em dashes. This punctuation mark has become widely recognized as a potential indicator of AI-generated text in recent times.

The Rise of the "ChatGPT Hyphen"

The frequent appearance of em dashes has been noted across various platforms. These include academic assignments, email correspondence, online discussions, professional networking sites like LinkedIn, and marketing materials.

Critics have suggested that the overuse of em dashes signifies a reliance on AI writing tools and a lack of original thought. Some have labeled it a sign of writers taking a shortcut.

A Pre-Existing Stylistic Choice

It's important to note that many writers utilized em dashes in their work prior to the emergence of large language models (LLMs). However, the consistent inclusion of the mark by chatbots contributed to its negative perception.

Even though not a definitive indicator of AI authorship, the em dash became an unwelcome addition to text for many readers.

OpenAI Addresses the Issue

For a period, ChatGPT users found it difficult to prevent the chatbot from employing em dashes, even when explicitly requested. This proved to be a frustrating limitation.

Sam Altman, OpenAI's CEO, has now confirmed that this issue has been resolved. He stated on X (formerly Twitter) that the chatbot will now adhere to instructions regarding em dash usage.

Custom Instructions Offer Control

OpenAI clarified on Threads that ChatGPT will demonstrate improved adherence to user preferences regarding em dashes. This is achieved through the use of custom instructions within the personalization settings.

The update doesn't eliminate em dashes entirely, but provides users with greater control over their frequency in generated text. Users can now actively manage this stylistic element.

This represents a “small-but-happy win” for users seeking more nuanced control over ChatGPT’s output.