ChatGPT Now Refuses to Mimic Famous Authors' Writing Styles, Citing Copyright

What Happened

ChatGPT has started refusing requests to write in the style of specific famous authors, according to testing by Ars Technica and confirmed independently by Engadget. The behavior applies to both living and dead authors, marking a departure from how the chatbot previously handled such prompts.

When Engadget asked ChatGPT to write a mystery story about a woman investigating her husband's disappearance "in the style of Agatha Christie," the chatbot declined: "Agatha Christie's works are still under copyright, so I can't provide text that closely imitates her distinctive style."

Notably, ChatGPT didn't stop there. It offered an alternative: "I can certainly write something with those characteristics, while remaining original." The bot then produced text carrying Christie-like qualities — atmospheric tension, a country-house setting, methodical clue-dropping — without directly citing itself as imitating her prose.

What Changed

A prior study by publication No Latency found that ChatGPT did not refuse similar requests involving deceased authors, suggesting OpenAI has updated its content policies or model guardrails recently. Neither Engadget nor Ars Technica could pinpoint an exact date for the change, and OpenAI has not issued a public statement confirming or explaining it.

The distinction OpenAI's system appears to be drawing is a narrow one: it will decline to "closely imitate" a named author's "distinctive style" but will still generate text sharing thematic or structural characteristics associated with that author. Critics and users testing the behavior have noted this makes the refusal feel more like a disclaimer than an actual content block, since the practical output — a Christie-flavored mystery — remains available under a different framing.

Why It Matters

OpenAI is currently facing multiple copyright lawsuits from authors, publishers, and media companies over its training data and outputs, including suits from the Authors Guild and individual novelists. According to Engadget's reporting, this behavioral shift is widely interpreted as a preemptive move to reduce legal exposure rather than a technical limitation — the underlying model clearly retains the capability to produce Christie-style text, since it does so anyway when the prompt is reframed.

This mirrors moves other AI companies have made around named-style requests for visual artists in image generation tools, where similar copyright pressure has led to comparable refuse-but-substitute behaviors.

What This Means

This is a policy-layer change, not a capability change. ChatGPT can still produce Christie-esque prose — it just won't do so if a user explicitly names her as the target. That distinction matters for legal risk more than for practical output, since the refusal is easily routed around by rephrasing the request. Expect other AI vendors facing similar authorship litigation to adopt comparable guardrails, and expect users to keep finding workarounds, since the underlying training data and stylistic capacity haven't been removed — only the explicit naming of the source has been restricted.