OpenAI addresses a notable concern regarding the overuse of em dashes in ChatGPT’s writing, a detail that has become a clear indicator of AI-generated text. Users highlighted that this stylistic quirk made it easier to identify ChatGPT’s output. To enhance personalization, OpenAI introduced a custom instructions feature allowing users to request minimal use of em dashes, resulting in a more natural writing style. This change aims to increase flexibility and trustworthiness in ChatGPT. Reactions have been diverse; some attribute the em dash prevalence to the training data, primarily books where such punctuation is common. Others suggest it emerged from models before integrating user-generated content. Notably, some users argue em dashes enhance text flow, pointing out their efficiency in maintaining sentence coherence. As OpenAI continues to refine ChatGPT, this focus on user control over tone and style reflects a commitment to improving the AI’s writing quality and adaptability.
Source link
