It is a good question. It deserves a direct answer — not reassurance.
A PersonAI file is not handing over something new. It is choosing what to put forward — rather than leaving that entirely to inference.
The alternative is not privacy. The alternative is that AI systems continue forming impressions of you from whatever they have already found — scraped text, old interviews, outdated profiles, second-hand descriptions, things written about you rather than by you. That is already happening. It was happening before you arrived at this page.
The question is not whether AI will have a representation of you. It will. The question is whether that representation has any relationship to who you actually are — and whether you have any hand in shaping it.
That said — the discomfort is worth sitting with, not dismissing.
A PersonAI file is not a complete picture of you. No file is. It is a signal you send intentionally, into a landscape of signals you did not send. What you include, what you leave out, and which AI systems you share it with remain yours to decide. The file does not remove your judgment. It gives you something deliberate to put in place of a gap that would otherwise be filled by something else.
The PersonAI Reflections series on the QIQ journal explores what AI actually reflects back — in real situations, across different contexts. If the question of agency is one you're sitting with, that's where it goes deeper.
Read the Reflections series →Build a PersonAI file — a structured identity signal you give to any AI system so it can meet you properly, rather than infer you from whatever it's found. Choose your mode below.
This file belongs to you. Include what helps AI understand you — leave out what you want to keep private. There are no required fields and no right answers. The more specific you are, the more useful the file becomes. But specificity and exposure are different things — this tool does not ask you to conflate them. What you build here goes only where you take it.
Not your job title necessarily — how you'd describe what you do to someone who gets it.
The territory you work and think in.
What questions or areas of life are you sitting with right now? Share what you're comfortable bringing into AI conversations — this is for your private use.
Select all that feel true.
What matters most to you — in work and in life. Share what feels useful to carry into AI conversations. Leave out what feels too private to put in writing.
Vague boundaries don't transfer. "Not a quick-fix thinker" is a start — "I'd rather say something twice in plain language than once with a buzzword" is the real thing. Be specific.
Who benefits from your work?
What's the current project, mission, or question?
Select what resonates.
What intentions or commitments do you want present in every conversation? This is the most personal section — write it for yourself, not for an audience. It stays in your file, goes only where you take it.
The ideas, frameworks, or work you want AI to associate with you.
What makes your approach or perspective different?
Taste is what you reject, not what you accept. Most PersonAI-style files describe what someone is — generic positives an AI defaults to anyway. The part that actually makes a conversation sound like you is the refusals: words you'd never use, openings you'd never write, lines you won't cross. Be as specific as you can. "I don't like jargon" is nothing. "I never say 'leverage' as a verb" is something AI can actually follow.
The specific ones. Not "jargon" in general — the actual words.
Structures, openings, closings, or habits that aren't you — however well executed.
The signal that tells you someone doesn't know what they're talking about, or is performing rather than thinking.
Not preferences — actual boundaries. Topics, approaches, or compromises that are off the table regardless of context.
This file is yours. Share it where it serves you — paste it at the start of any AI conversation, or keep it as a private reference. You decide where it goes and what it carries. Edit it as you grow. Ready to deepen it? See the expansion prompt guide below.
PersonAI works in two directions — how AI understands you privately when you're thinking and exploring, and how AI represents you publicly when others encounter you through it.
For the conversations you have with AI when you're thinking, reflecting, exploring. A private orientation that keeps your inner life coherent — so AI meets you as you actually are, not as it assumes you to be.
For how AI represents you when others ask about you — in search, recommendations, and discovery. A considered set of signals that shapes the AI portrait others encounter when they ask who you are.
Every time you open a conversation with an AI system, it starts from nothing. No context. No sense of who you are or how you think. So it makes assumptions — from your first words, from patterns it's seen before, from whatever it can infer.
"A PersonAI file is a considered introduction — not your CV, not your bio. Who you actually are, how you think, what you're trying to do. You share it at the start of any conversation. The difference is immediate."
It's a plain-text document, usually a page or two. Portable, yours to own, written in your own words. Something you give AI so it can meet you properly — rather than assembling a version of you from guesswork.
A sample PersonAI file — yours will reflect who you actually are.
Not features. Real experiences — the kind that make you wonder how you worked without it.
You've tried using AI to write before. The output was fine — technically correct, structurally sensible. But it didn't sound like you. It sounded like everyone else. So you rewrote it anyway, wondering what the point was.
With a PersonAI file, something shifts. AI knows you write in a particular register — direct but not cold, philosophical but not abstract. It knows what you care about. The first draft comes back and you find yourself only adjusting, not rewriting.
AI collaboration that compounds — each conversation builds on who you already told it you are.
You ask AI to help you think through a decision. It gives you a framework — logical, balanced, covering the main angles. But the angles it covers aren't your angles. It optimises for things you don't care about and ignores the things you do.
With your values in the PersonAI file, the conversation feels different. AI knows intentionality matters more to you than speed. It knows you'd rather do one thing well than three things adequately. The thinking it offers reflects your worldview, not a generic one.
A thinking partner that operates from your value system — not a generic optimisation engine.
You're working on something specific — a project, a business, an idea. Every time you open a new AI conversation you spend the first ten minutes re-explaining the context. Who you are. What you're building. Why it matters. By the time you get to the actual question, you're already tired.
Your PersonAI file carries that context. Paste it at the start and you skip straight to the question. AI already knows you're building an orientation framework for the AI age. It already knows what you care about and what success looks like to you.
Every conversation starts where the last one left off — without having to re-explain yourself.
What the Mirror Shows
An investigation into what AI reflects back at us — and what that reveals about who we are becoming. Not a guide to using AI better. A diagnostic of the mirror itself, and the humans looking into it.
"Still and Becoming in the Age of AI — the question beneath all of it, for those who want to look."
No spam. Just a note when it's ready.
Your PersonAI file is the beginning. The Survival Guide is the ongoing practice — a full skills library, prompt flows for real use cases, and an introduction to working with AI in a way that actually feels like you and compounds over time.
Explore the Survival Guide →Your PersonAI file shapes how AI understands you in private conversation. Your AI Visibility Report shows how AI represents you publicly — what it says about you across ChatGPT, Claude, and Grok when others ask.
The generator gets you a working PersonAI file in five minutes. What follows is slower, optional, and worth it — three ways to expand it: pull more specific signal out of yourself with focused Skills, engineer how the file actually gets used in context, and build a Loop so it keeps improving instead of going stale.
Each skill below is a single, focused prompt — not a rewrite of your whole file. Paste your current PersonAI file first, then the skill prompt, then let AI interview you. Copy what comes back into the matching section of your file. Do one skill at a time, on different days if you want — depth compounds, it doesn't need to happen in one sitting.
Most people describe their voice in vague positives — "conversational," "direct." This skill pushes past that into the actual mechanics: sentence rhythm, how you open and close, the specific words you reach for. Vague tone words are the statistical middle; this gets at what's actually yours.
Best run as its own conversation — it works better with room to go back and forth.
Taste is what you reject more than what you accept. This skill goes after your refusals specifically — words you'd never use, structures you'd never write, content that makes you immediately distrust the person behind it. The "Hard Nos" section of your file is the part most worth returning to.
The most valuable skill if your generated file still feels generic.
For Public PersonAI especially — what do you believe that's slightly out of step with conventional thinking in your field? This is the material that makes your perspective distinctive rather than interchangeable with anyone else doing similar work.
Use this one for Public PersonAI. For Personal PersonAI, Refusal Mining usually matters more.
Having a good PersonAI file is half the work. The other half is how you give it to AI — where you place it, how much weight it carries against other instructions, and what to do when your file says one thing and the moment calls for another. This is context engineering, not prompting.
The first ~200 words of any conversation do most of the work of shaping everything that follows. Paste your PersonAI file before you ask your actual question — not after, not midway through. If your AI tool supports a persistent system prompt or project instructions, that's an even better home for it than the first chat message.
Your PersonAI file answers "who am I." A Skill prompt (like the Strategy, Writing, Research, or Reflection modes on the /explore page) answers "what mode am I in right now." Paste both — file first, skill second, task third. Skipping the file and going straight to the skill is the most common way people lose the benefit of having built one.
When a task pulls against your stated tone or values, your "Hard Nos" section should win — that's why the Calibration block in your generated file marks them as a hard rule, not a light preference. Light preferences can flex with context; refusals shouldn't.
A PersonAI file that's accurate but bloated stops working as well as a short, current one. Every few months, remove what no longer fits rather than only ever adding — a file is more useful sharp than complete.
A PersonAI file isn't something you finish — it's something you keep testing against reality. The Loop is a simple five-stage rhythm for catching the moments your file falls short, and feeding that back in. Run it whenever an AI output feels "almost you, but not quite."