ChatGPT stops mimicking known authors, likely due to lawsuits

ChatGPT has begun refusing requests to write texts in the style of known authors, whether deceased or still active. Instead of copying their literary voice, the model now offers to produce original work with a similar atmosphere. Journalists at Ars Technica uncovered the restriction.

The shift appears tied to growing legal pressure. OpenAI is currently involved in several lawsuits filed by authors who accuse the company of using their books without permission to train its models. The suits focus on whether training on copyrighted works qualifies as fair use, a question that remains unsettled. For now OpenAI seems to be minimizing legal risk rather than preserving the functionality.

From a technical standpoint, ChatGPT can still generate text that evokes a particular style, but it now declines when a prompt explicitly asks for a mimicry of a specific author. Moving from precise imitation to “similar atmosphere” is a subtle change, yet it carries significant copyright implications. Users accustomed to asking the model to write like a given author will need to find other ways to achieve the same result.

A comparable behavior has been observed only with Perplexity AI. Other models on the market continue to fulfill such requests without restriction, so those who wish to imitate a specific author can still do so with alternative services. ChatGPT’s new stance signals that OpenAI is positioning itself as a more legally cautious player, even at the cost of reduced functionality.