The human side of AI

AI is changing how you think.Start noticing.

A private calibration across the eight human capacities AI quietly taxes — attention, agency, trust, connection, recovery, ambiguity, relationship, judgment. Eight statements. Nothing leaves your browser.

24questions
08chapters
0accounts
Ready·Chapter 01 — Arrival0:00 / 4:57

Hello?

Recorded session. Two voices, one microphone, no edit.08 chapters · synced transcript

The voice below is a recorded 4:57 conversation, not an AI answering you live. Turn your sound on.

Scroll to the signals

The important AI question is no longer what it can do. It is what it is doing to you.

Hear the conversation

01 — What we measure

Eight capacities.
None of them on a dashboard.

AI adoption is measured in tokens, seats and hours saved. These are the capacities that actually pay for it, and none of them show up in a usage report.

Trust04 / 08

Signal 04

Do you check what matters?

Checking is unglamorous, slow, and the only thing standing between a useful draft and a confident mistake.

Cycling

02 — Calibration

Start with
a better question.

One real question from the assessment, with a reflection that changes depending on your answer. No account, no email, no score to chase.

One question from the calibration · 24 in full

When AI offers an answer, how often do you pause to check it?

How this works

Answer it for real. You will get a reflection, not a score.

Nothing is stored, nothing is sent. The full calibration is eight short chapters and runs entirely in this browser.

03 — For organizations

Adopt AI.
Keep your people strong.

Individuals see their own picture. Leaders see the pattern. Nobody sees a person, and the room cannot report on fewer than twelve people.

Private by design, not by policy.

The calibration is answered in the browser and belongs to the person who answered it. Aggregation happens on device counts, never on individual rows — so a manager cannot ask for “just one person’s result”, because the system never has one.

01A private calibrationEmployees answer without their responses appearing in an employer dashboard.
02Anonymous aggregationGroup-level signals only, suppressed below twelve respondents.
03A measured interventionLiteracy, focus protection, verification habits, and a re-assessment plan.
The organizational model

Pattern room · team of 40 · last 14 days

Simulated

Leaders see the shape of the room, never a person. Recovery at 33 is a work-design finding, not an individual one.

No individual answers · n ≥ 12 to display

04 — Evidence

Evidence before
certainty.

There is real research here, and there is a great deal of guessing dressed as research. We label the difference on every claim we make.

Established

Known patterns worth applying

Automation bias, digital wellbeing, workplace burnout: substantial bodies of work that already tell us what to do.

Emerging

Signals worth studying

Early research on generative AI, cognitive engagement, reliance and social behaviour. Suggestive, not settled.

Open

Hypotheses, not headlines

Questions that need longitudinal study before anyone, including us, should make a strong claim about them.

Every claim on this site is labelled by confidence, and no citation is invented

05 — Human support

When it matters,
speak to a person.

A wellbeing product that only offers software is not a wellbeing product. Humenra is designing a route out of the screen — starting with one confidential conversation with someone qualified.

Explore human support
01 — GuideA confidential first conversation to work out what kind of support fits.
02 — ProfessionalA referral pathway to licensed counsellors and psychologists, where appropriate.
03 — CommunityFacilitated peer groups and offline programmes built for actual company.

Ready when you are

Eight minutes. One honest look.

Eight chapters, a reflection after each one, and a balance card you can keep. Your answers stay in this browser — closing the tab closes the session.