AI Diary Study

A semester with your AI,
one entry at a time.

Ten minutes of honest writing each week about you and the AI you actually use. You will watch your own story develop; researchers will learn how these relationships form.

Enter an email you check regularly; any address works. You sign in with a one-click link, there is no password, and after the first time this device stays signed in.

New here? See the full walkthrough with screenshots.

How a week goes

  1. 1

    One link, once a week

    A sign-in link arrives in your email. No password; two clicks and you are writing.

  2. 2

    Ten minutes of honest writing

    A quick check-in, a prompt, and room to write. Drafts save as you type.

  3. 3

    Something back, every time

    The moment you submit, you see your own writing patterns. When the week closes, you see how they sat alongside the class. Never anyone's words; just the shape of the week.

The feedback, at a glance

The way you wrote this week

Sample, with made-up numbers

Positive emotion "happy", "enjoy", "hope"

you 3.4% · class 2.6%

up 1.3 from your last entry

Negative emotion "worried", "stressed", "alone"

you 1.1% · class 1.9%

down 0.1 from your last entry

Self-focus "I", "me", "my"

you 7.8% · class 6.2%

up 0.3 from your last entry

Thinking words "because", "realize", "maybe"

you 5.6% · class 4.4%

up 1.6 from your last entry

After every entry, a mirror held up to your own language: how warm it ran, how much of it was about you, how hard it was working through something, next to your own last entry. Once each week closes, the class median joins the picture, and your home page keeps every week's results for the whole semester. Always the class in aggregate, never another person's writing.

What the writing becomes

Everyone is suddenly living with AI, and almost nobody is keeping a record of what that is actually like. Week by week, your entries build one: a semester-long account of a person and a new technology figuring each other out, written while it happened rather than remembered afterward.

On the research side, the writing is analyzed as language. Dictionary-based software running on our own computers counts things like emotion words, pronouns, and style, and we study how those patterns move across the semester. Your words are never sent to an AI service, results are only ever reported in aggregate, and your entries carry a random code rather than your name.