OpenAI has announced that ChatGPT will now give users the option to eliminate em dashes from its generated text. The excessive use of em dashes, a common feature in AI-generated content, has been a point of contention in recent months across various platforms such as school papers, emails, customer service chats, and online forums. Critics have accused writers of laziness and overreliance on AI chatbots, citing the prevalence of the so-called “ChatGPT hyphen” as evidence.
While some writers have defended the use of em dashes as a longstanding part of their writing style, the inability of chatbots to refrain from using them has made them a target of scrutiny. OpenAI struggled to address this issue, with users finding it challenging to prevent ChatGPT from incorporating em dashes even when explicitly instructed not to do so.
OpenAI CEO Sam Altman recently announced a solution to this problem, stating that ChatGPT will now adhere to user preferences regarding em dashes. Altman shared the update on Twitter, describing it as a “small-but-happy win” for the platform. Users can now customize their instructions in ChatGPT’s personalization settings to control the frequency of em dashes in the generated text.
In a post on Threads, OpenAI elaborated on this update, emphasizing that users can now instruct ChatGPT not to use em dashes, thereby reducing their presence in the output. While em dashes may still appear in generated content by default, users now have more control over their usage, addressing concerns about the overabundance of this punctuation mark in AI-generated text.
