How we use AI
We use AI, and we are not apologetic about it. What we will not do is let it decide anything about you.
Where we stand
It let one person move a stone that ten could not, and what got built afterwards was not the same as what came before. Nobody argues the builder was diminished by it.
AI is a tool of that kind, and knowing how to use it is fast becoming part of the job for anyone who works with their head. We would rather say that plainly than pretend otherwise — it would be a strange position for a product whose most useful feature reads your CV so you do not have to retype it.
So the question is not whether to use it. It is what you point it at. Doing work for a person and making decisions about them are not the same thing, and almost everything that has gone wrong with AI at work is the second one wearing the clothes of the first. A CV sifted by a machine that never read it. A score nobody can see or argue with. A rejection with no person behind it.
That is the line. On this side of it, a tool that saves you an afternoon. On the other, a system that decides your afternoon for you.
In the product
All four are things you ask for. None of them runs on its own, and none writes anything you have not seen.
You upload a document, a model turns it into a record, and you check it before anything is saved. It is doing the typing you would otherwise do yourself — nothing is kept from the document beyond what ends up on your record.
Only if you tick the box. It is marked as AI-written where it appears, and you can rewrite or delete it. Off by default, because most people would rather write their own.
It reads a draft and tells you what it thinks. It cannot touch your post — not by policy but by construction: the code that runs the assistant has no route to your writing. Every change is one you make.
Ask for a version of your record aimed at a particular audience, and a model chooses what to lead with and what to leave out. It never invents an employer, a job title, a date or a number — those are copied from your own record. A fabricated job is not something we catch; it cannot be produced.
The safeguards
Not your record, not your writing, not your CV. Not by us, and not by anyone we send data to.
Prompts and responses are not stored or retained for any purpose by the models we call. They process the request and it is gone.
Nothing here scores you, ranks you, sorts you or judges you. There is no model deciding who sees your record or whether you are worth an employer's time. When employers arrive, the rule is the same — a person reads your application.
Every model output lands in front of you first. Nothing is written to your record, published, or sent anywhere on a model's say-so.
How you can tell
Anyone can write a page like this. What is worth more is when the software cannot do the thing even if we changed our minds.
When a model builds you a variant of your record, it is not given the ability to write an employer, a title or a date. Those are copied from rows you already own. It is not that we check for invented jobs — there is no path by which one could appear.
The writing assistant is the same. It has no write access to your posts, so it cannot change your words whether it means to or not. It can only tell you what it thinks.
And when a machine has read something of yours, we say so on the screen where it happened, not in a policy. Your record shows that it was built from a CV by AI, and asks you to check it, until you say you have.
If you think any of this is wrong, or we have drifted from it, tell us: support@unlinked.pro.