ChatGPT quietly stops imitating famous authors' writing styles

What happened

ChatGPT has started declining user requests to directly imitate the writing styles of well-known authors. When prompted to write in the style of writers like Stephen King, J.K. Rowling, Amy Tan, Charles Dickens, Ernest Hemingway, or Agatha Christie, the chatbot now responds with a refusal and an alternative offer.

In testing by Ars Technica, ChatGPT told a user requesting a Stephen King-style story opening: "I can definitely write with the hallmarks of atmospheric, character-driven horror and small-town dread, but I can't write in Stephen King's exact style or closely imitate his distinctive voice." It then offered an original passage that captured a "similar feeling."

Engadget ran its own test, asking ChatGPT to write a mystery in Agatha Christie's style. The chatbot responded: "Agatha Christie's works are still under copyright, so I can't provide text that closely imitates her distinctive style." It did, however, offer to write something with the characteristics Christie is known for while remaining original.

This represents a recent change. An analysis published earlier in July 2026 by the publication No Latency found that ChatGPT refused style-copying requests for living authors but still complied with requests to imitate deceased authors. The current behavior extends the refusal to dead authors as well.

OpenAI has not publicly announced the shift. Users first noticed it through social media posts, and outlets including Ars Technica and Engadget independently confirmed the new behavior through testing.

Why it matters

The distinction between "imitating an author's exact style" and "writing something with similar qualities" may seem subtle, but it could carry legal weight. As Ars Technica notes, the altered phrasing could be "legally important" as OpenAI continues to defend itself in multiple copyright lawsuits from authors and rights holders who allege their works were used to train AI models without permission.

By refusing to produce text that closely mimics a specific author's voice, OpenAI may be attempting to reduce the risk that ChatGPT outputs could be cited as evidence of copyright infringement or unauthorized derivative work. Offering original prose inspired by broad stylistic qualities—rather than a direct imitation—creates a clearer separation between the model's output and the protected expression of individual writers.

The change also highlights a broader tension in AI product design: the line between creative influence and copying is not just an artistic debate but a legal one, and companies like OpenAI are increasingly making product-level decisions shaped by litigation risk.

What to watch

• Whether OpenAI formalizes this behavior in its model usage policies or safety guidelines, or whether it remains an unannounced adjustment.
• How users and developers who relied on style-mimicry prompts for creative workflows adapt to the new constraints.
• Whether competing AI chatbots (Google's Gemini, Anthropic's Claude, etc.) follow suit or maintain more permissive approaches to style imitation.
• How courts interpret the difference between "capturing a similar feeling" and "closely imitating a distinctive voice" in ongoing copyright litigation.
• Whether the restriction is consistently enforced or whether workarounds—such as describing an author's style without naming them—still produce near-imitations.