OpenAI Quietly Bans Author Mimicry in ChatGPT, but a Simple Loophole Remains

OpenAI has quietly altered ChatGPT so that it now rejects direct requests to mimic named authors, blocking prompts that ask for writing 'in the style of' famous writers while still allowing users to generate original work using similar techniques, tones and structures.

A request that once produced a familiar creative experiment with ChatGPT now leads to a refusal. Ask the chatbot to write like Stephen King, J.K. Rowling or Ernest Hemingway, and it will no longer attempt to reproduce their distinctive voices. But the underlying ability to generate similar moods, structures and storytelling techniques remains available.

OpenAI has quietly adjusted ChatGPT so that prompts asking for an exact imitation of a named author are rejected, while broader requests describing tone, atmosphere and narrative style continue to work. The change alters how users ask for creative output rather than removing the model's understanding of literary styles.

ChatGPT Rejects Requests To Copy Famous Writers

ChatGPT users requesting text 'in the style of' a specific author are now receiving responses that refuse direct imitation.

A prompt asking for a story opening in the style of Stephen King, the American horror novelist known for psychological suspense and small-town settings, produces a response explaining that ChatGPT cannot write in the author's exact voice. It instead offers to create an original piece using similar elements, such as tension, fear and an unsettling atmosphere.

The same response pattern has appeared in requests involving other writers, including J.K. Rowling, the British author behind the Harry Potter series, Agatha Christie, the English mystery novelist, and Ernest Hemingway, the American writer known for concise prose.

The restriction appears to apply to both living and deceased authors.

OpenAI has not publicly announced the change or released a formal update explaining when the behaviour was introduced

The Loophole Is in the Prompt, Not the Model

The update does not prevent ChatGPT from producing writing with characteristics associated with particular authors.

Users can still describe the elements they want without naming a writer. A request for a horror story set in a small American town, with ordinary characters facing disturbing events and a growing sense of dread, can produce a similar atmosphere without directly asking for Stephen King's voice.

A mystery prompt focused on hidden motives, complex clues and unexpected reveals can generate a story built around familiar detective fiction techniques without requesting Agatha Christie's exact style.

The distinction is between asking the model to replicate a specific author's voice and asking it to apply broader writing methods.

ChatGPT can still analyse how writers use language, pacing, structure and narrative devices. The change is that an author's name is no longer a direct shortcut for generating that style.

Copyright Disputes Put AI Writing Practices Under Scrutiny

The change comes as artificial intelligence companies face ongoing legal disputes over copyright, training data and AI-generated content.

OpenAI, the company behind ChatGPT, has been sued by groups of authors who allege that copyrighted works were used without permission to train AI systems.

Writers including George R. R. Martin, the fantasy author behind A Song of Ice and Fire, and John Grisham, the legal thriller novelist, are among authors involved in copyright cases against AI companies.

Publishers have also brought legal action over claims involving the use of copyrighted material in AI training and the output produced by generative AI systems.

US copyright law generally protects specific expression rather than broad ideas or writing styles. However, AI-related disputes have examined whether generated content can become sufficiently similar to protected works.

OpenAI has previously introduced safeguards designed to prevent ChatGPT from producing copyrighted passages or near-verbatim reproductions of existing material.

AI Companies Draw Different Lines on Style Imitation

Technology companies have taken different approaches to requests involving famous writers.

Some AI systems continue to allow users to request writing inspired by specific authors, while others restrict direct style imitation. The differences reflect a wider debate over how generative AI tools should handle creative influence.

AI models are capable of identifying patterns in text, including sentence structure, vocabulary, pacing and tone. The debate centres on where the boundary sits between using those elements as inspiration and reproducing a recognisable creative identity.

OpenAI Changes the Route, Not the Destination

The ChatGPT update changes how users access style-based writing prompts, but it does not remove the model's ability to generate work with specific characteristics.

A direct request to imitate a named author is now more likely to trigger a refusal, while a prompt describing desired qualities remains available.