Responsible AI
AI that reads your book — not one that learns from it.
How we use AI to give you feedback — and where the line sits between your work and a machine’s training data.
Our proprietary critique system
Critique Partner is our own system, engineered for long-form fiction. It reads with your whole book in mind — not just the page in front of it.
A large language model is the raw material almost every modern AI product starts from — the shared foundation, not the difference. What makes Critique Partner is the system we built on top of it. That system is the product, and it’s ours.
We never train AI on your writing
This is the promise that matters most. Your novel is never used to train, fine-tune, or improve any AI model — not ours, and not our providers’.
We send chapters to established providers — such as Anthropic and OpenAI — through their commercial APIs, not consumer chatbots; under those terms they don’t train on it. We name them so you can read their policies yourself.
How a critique is made
Each partner is part of a proprietary system we built for long-form fiction, with leading AI models engineered into its core. The judgment is ours; the models supply the fluency. What we send and keep is spelled out on the Privacy page.
You keep your authorship and copyright
Your words are yours. Using Critique Partner transfers no ownership — feedback is a suggestion to take or leave, and the book that results is yours.
AI gets things wrong — you’re the editor
AI feedback is a thoughtful second read, not a verdict. It can misread your intent or miss what a human would catch — weigh it, argue with it, keep what helps. The final call is always yours.
This describes how Critique Partner uses AI today — we won’t state a protection here until it’s actually in place.