Show Quick Read Summary is AI-generated, newsroom-reviewed

New Delhi: OpenAI is bringing a solution to a problem that existed but was rarely spoken about in public - the inclusion of too many em dashes in ChatGPT’s writing. What seemed like a minor stylistic quirk has now been acknowledged by the company as a clear marker of AI-generated text today.

Many ChatGPT users have pointed out that the repeated use of em dashes has become a giveaway. Readers could often identify ChatGPT’s output simply by noticing the punctuation pattern.

OpenAI has now introduced a fix through its custom instructions feature, allowing users to tell ChatGPT not to use em dashes. The model adapts quickly when given this directive, producing text that feels more natural and less formulaic.

This change is part o

See Full Page