This guide is about a different approach: a genuine, verifiable record of how a piece of work came to be, built as a byproduct of simply doing the work. It is written for writers navigating authorship in the age of AI.
How we think about this - Throughout, this is about evidence and reassurance — never proof, never certification, and never detection. A verifiable record can corroborate how work developed; it cannot read your mind or your manuscript. And a record can only ever speak for a writer, never against one: a thin record or an absent one proves nothing, and must never be treated as if it did.
In this guide
- The tools that watch you type — and the record that doesn't
- Backing your "no AI" declaration for awards
- If your work is ever questioned
- What a verifiable record is — and isn't
- Why we never read your manuscript
- Provenance for co-writers and ghostwriters
The tools that watch you type — and the record that never reads your words
Most tools promising to prove human authorship do it by surveillance — recording your keystrokes, storing your drafts, watching you write. There is another way. A record built from the business of managing your work — the stages a project moves through, the timeline of its development — can evidence that a real project took real shape over real time, without any tool ever reading a word of your manuscript.
The difference matters more than it first appears. One approach treats the writer as a suspect to be monitored; the other treats the record as a natural byproduct of working — held by the writer, revealing process without ever exposing prose. If a tool has to watch you type to trust you, it has already decided you are guilty.
Backing your "no AI" declaration for awards
More and more New Zealand and Australian awards now ask entrants to declare that their work is their own, unaided by AI. A declaration is only as strong as the trust behind it — and a verifiable record of how the work developed lets you stand behind that declaration with something an awards trust can actually check, without ever submitting your manuscript to anyone.
This is not about proving a negative or beating a detector — a record can do neither, and should never claim to. It is about being able to show, if you are ever asked, the genuine shape of how a piece came to be: a quiet, honest answer to a question that is only going to be asked more often.
If your work is ever questioned: process, not detectors
If a piece of your writing is ever wrongly suspected of being machine-generated, the thing that clears an honest writer is not a detector score — detectors are unreliable, and increasingly distrusted — but evidence of genuine process: the drafts, the timeline, the development that generated text simply does not have. In a widely reported 2026 literary case, accused writers were cleared exactly this way: by their working record, not by a machine.
A verifiable development record gives you that evidence in advance, quietly accumulated, should you ever need it. And to be absolutely clear: a thin record is never held against you. Its absence proves nothing; its presence simply reassures. The record exists to help the honest writer, never to accuse anyone.
What a verifiable writing record is — and isn't
A verifiable writing record is corroborating evidence of process — a timestamped, tamper-evident account of how a project developed over time. It is not proof of authorship, not an AI detector, and not a certificate. It cannot read your mind or your manuscript, and it does not try to. Being clear about what it is not is precisely what makes what it is trustworthy.
Honesty about the limits is the whole point. A record that claimed to prove more than it can would deserve exactly the scepticism detectors have earned. This one corroborates; it does not adjudicate — and that modesty is a feature, not a shortfall.
Why we never read your manuscript
The record Writion keeps is built entirely from business-layer information — the stages a project moves through, when things happened, the shape of the work over time. It never reads, stores or analyses your actual words. "Metadata, not manuscript" is not a feature we added; it is the line the whole thing is built around.
This matters more in the AI age, not less. The tools asking you to prove your humanity should not require you to hand over your work to do it. A record that reveals process without exposing prose respects both things at once: the genuine need for reassurance, and the writer's plain right to their own words.
Provenance for co-writers and ghostwriters
Collaboration raises its own questions — who did what, and when. A process record lets each collaborator evidence their own contribution and timeline, without exposing the text itself. For ghostwriters especially, it offers a way to hold a professional record of the work done, on terms that protect everyone's confidentiality.
It records activity and timeline, not claims about specific sentences — process, not the attribution of every line. Within those honest limits, it brings a quiet clarity to collaborations that often have very little of it.
This is the part of the AI moment Writion was built for — not to detect, not to certify, not to watch you write, but to let the honest record of your work exist as a byproduct of doing it. The record is yours. Your words stay yours. And if the question ever comes, you have something genuine to stand behind.
About Writion. Writion is a writing business management platform built for professional writers in New Zealand and Australia — the one place where both the commercial work and the creative work live. Founded by author Gaye Miller. Writion records only business-layer information; it never reads or stores a writer's manuscript.