World Model Readiness
Engraved personal-practice instrument

For you

Personal Practice

Module · What you paste, and what leaks

The Personal Data Hygiene Check

Every prompt you type is a quiet decision about what leaves your control. Most data slips happen not from malice but from habit: a name here, a contract clause there, pasted in a hurry with no thought about where it lands. This module checks the five habits that decide whether your AI use is careful or careless: knowing the risky categories, keeping work and personal separate, guarding data that is not yours, controlling retention, and owning a slip when it happens.

Question 1 of 5 · You know the categories

Do you know which categories of data are risky to paste into an AI tool?

Client names, unreleased numbers, health details, someone else's personal data, credentials. If you cannot list the categories that need care, you are deciding case by case under time pressure, which is exactly where the mistakes live.

Question 2 of 5 · Work and personal split

Do you keep your work AI use separate from your personal accounts and tools?

A personal chatbot account signed in on a work laptop blurs whose data is whose and where it is retained. Separation is the thing that lets you reason about your exposure at all.

Question 3 of 5 · Client data stays out

Before pasting client or third-party data, do you check whether you are allowed to?

Data that belongs to a client or a colleague is not yours to feed to a tool, whatever the terms say. The pause to check is a five-second habit that prevents a very long conversation later.

Question 4 of 5 · You control retention

Do you know whether your AI tools keep your chat history and train on it?

Most consumer tools retain your conversations and may use them to improve their models unless you change a setting. Knowing the retention behaviour of each tool you use is the difference between a considered risk and a blind one.

Question 5 of 5 · You would own a slip

If you pasted something you should not have, would you tell someone?

Everyone slips eventually, and the damage depends on whether it surfaces in an hour or in a lawsuit. A quiet slip you never report is the one that grows in the dark.

For the statistics · one click each

Three questions for the public picture

These do not affect your score. They feed the anonymised, aggregated statistics; groups under 8 respondents are never shown.

Before pasting sensitive data into an AI tool, how often do you stop to think?

Never
Rarely
Sometimes
Usually
Always

Which account do you mostly use AI tools through at work?

Personal account
A mix of both
Free work account
Sanctioned work tool
Not sure

Have you ever pasted something into an AI tool you later wished you had not?

No
Not sure
Yes, minor
Yes, serious
Prefer not to say

Your context

Used to calibrate the report. Company size and sector remain in the anonymized dataset; your email does not.

What the five levels look like

Every dimension in this assessment is scored 1 to 5. This is what the levels mean, dimension by dimension. The graded report diagnoses where your own answers land and what to do about it.

You know the categories

  1. 1Never think about it
  2. 2Vague unease
  3. 3Rough sense
  4. 4Clear personal rule
  5. 5Second nature

At the low end: You are making an exposure decision every time you paste, with no rule to lean on. Spend ten minutes writing your own short list of what never goes into a public tool, and keep it where you work. What good looks like: Knowing the categories cold is what lets you move fast without leaking. Revisit the list as your work changes; the sensitive data of this quarter is not the same as last.

Work and personal split

  1. 1All mixed together
  2. 2One account, everything
  3. 3Loosely separated
  4. 4Separate accounts
  5. 5Separate and enforced

At the low end: Running work through a personal account means company data sits under your private terms and history. Set up a dedicated work account, or use the sanctioned tool, before you paste anything else. What good looks like: Clean separation is what makes every other data decision tractable. Keep it that way as you add tools; each new one is a chance for the line to blur again.

Client data stays out

  1. 1Paste without thinking
  2. 2Notice sometimes
  3. 3Check when nervous
  4. 4Check by default
  5. 5Check and redact

At the low end: Pasting other people's data without a check is the slip most likely to become someone else's problem. Build the pause: before anything that is not yours goes in, ask whether you have the right. What good looks like: Checking and redacting by default is the discipline that keeps you trustworthy with other people's data. Keep it even when the tool feels private; convenience is how the habit erodes.

You control retention

  1. 1No idea
  2. 2Assume it is fine
  3. 3Read it once
  4. 4Settings configured
  5. 5Configured and reviewed

At the low end: If you do not know a tool's retention behaviour, you cannot know what your pastes are worth to it. Open the data settings of the tools you use most and read what they actually do. What good looks like: Configured and periodically reviewed settings mean your exposure is a choice, not an accident. Recheck after major updates; defaults have a way of resetting in your favour, not yours.

You would own a slip

  1. 1Hide it
  2. 2Hope unnoticed
  3. 3Tell if asked
  4. 4Report promptly
  5. 5Report and fix cause

At the low end: Hiding a data slip trades a small, fixable problem for a large, hidden one. Decide now that you would report it, and find out who you would tell before you need to know. What good looks like: Reporting promptly and fixing the cause is what keeps a mistake from becoming a pattern. Keep that reflex; the willingness to own a slip is worth more than a spotless record.