OpenAI's AI assistant for writing, ideas, and images — which, since late July 2026, refuses direct requests to write "in the style of" a specific named author.
Last verified · 2026-08-09 · by Moe Ameen
ChatGPT is OpenAI's conversational AI assistant — the product that put a general-purpose model in front of hundreds of millions of people. You type a request and it drafts, rewrites, summarizes, brainstorms, reasons through a problem, and, through the GPT-Image models, generates pictures. For creators it has become a default first stop: a place to spin up hooks, captions, outlines, scripts, and headline variations in one chat window. On paid plans it runs OpenAI's frontier GPT-5.6 family; the free tier defaults to the older GPT-5.5 Instant.
What changed for content people is a quiet policy shift. Starting in late July 2026, ChatGPT began refusing direct prompts that ask it to write "in the style of" a specific named author. Ask it to sound like Agatha Christie, Hemingway, or a living novelist and it now declines, saying it can't closely imitate that writer's distinctive voice — then offers to write something with the same tone, atmosphere, and techniques while keeping the text original. The block started with living authors and extended to long-dead ones soon after. OpenAI never formally announced it. The move is widely read as a hedge against the copyright lawsuits OpenAI is defending, and it lands in the same period as Anthropic's reported $1.5 billion author settlement.
The important nuance: the capability itself is intact. The refusal is on the direct ask, not on the output. Describe the qualities you want — short declarative sentences, dry wit, a specific rhythm — and ChatGPT still produces them. And the behavior is tool-specific: reports indicate Gemini has complied with named-style requests directly, while Claude and Copilot answer with disclaimers, so a prompt that fails in ChatGPT may not fail elsewhere. See our news write-up at /news/chatgpt-blocks-author-style-imitation for the full timeline.
Honest framing for creators: ChatGPT is a superb drafting and thinking layer, but it is a single chat window. It writes one reply at a time, does not enforce a consistent brand voice across a week of outputs, generates no video, builds no brand-exact carousels, and cannot schedule or publish anything to a social platform. What you leave a session with is a draft, not a distributed content set.
The style-imitation block is a useful prompt to rethink the whole workflow. If your process quietly depended on "write this caption in the style of [famous creator]," ChatGPT just took that away — and the durable fix was never borrowing someone else's voice, it's owning a distinctive one of your own that no model will ever refuse to write in, because it's yours. That is precisely what Kompozy's Persona Brief encodes. You define tone, vocabulary, audience, and banned phrases once, and every output the engine produces is held to it. No "in the style of" prompt, no refusal, no drift between posts.
Concretely, ChatGPT is the fast idea layer and Kompozy is the production engine that turns those ideas into a published week. Draft a rough angle or a messy outline in ChatGPT, paste it into Kompozy Quick Ingest, and one source fans out into up to 18 formats — Persona Shorts and HeyGen avatar video, Clipped Shorts from long footage, brand-exact Carousel Posts via HyperFrames, Quote Graphics, Photo Posts, native Text Posts, a Blog Article, and an Email Newsletter — all governed by the same Persona Brief so a batch reads as one brand. Then Kompozy schedules and publishes the set across the eight primary social platforms plus blog and email, with a per-post review pipeline and Autopilot. ChatGPT hands you a paragraph; Kompozy hands you the campaign, in a voice the model can't refuse to use.
It stopped honoring the direct request. Since late July 2026, ChatGPT declines prompts that ask it to write "in the style of" a specific named author, saying it can't closely imitate that writer's distinctive voice. It still offers to write with the same tone and techniques while keeping the text original, so the underlying capability is intact — only the direct ask is blocked.
OpenAI did not formally announce it, but it comes as the company defends multiple copyright lawsuits from authors and publishers. "Write in the style of [author]" is exactly the kind of output cited in those disputes, so refusing the direct ask reduces the most quotable examples of alleged mimicry. It landed in the same window as Anthropic's reported $1.5 billion author settlement.
Yes — describe the qualities you want instead of naming the author. Short declarative sentences, a dry tone, a specific rhythm, and similar techniques all still work; the refusal is triggered by naming a specific writer, not by the style itself. The behavior is also tool-specific, so other assistants may respond differently.
No. ChatGPT drafts text and generates images in a chat window, but it has no scheduler, no per-platform reframing, no brand-voice governance across outputs, and no publishing layer. To turn a ChatGPT draft into captioned video, carousels, a blog, and scheduled posts across platforms, pair it with a content engine like Kompozy.
ChatGPT has a free tier plus paid individual plans — Go at roughly $8/month, Plus at $20/month, and Pro up to $200/month — with per-seat Business and custom Enterprise options; confirm current pricing on openai.com. Kompozy, by contrast, is credit-based (Starter $99/month, Pro $299/month) and covers generation across 18 formats plus multi-platform publishing, not just drafting.