AI MMI Trainer — Marking Explained

How Scoring Works

A simple, honest look at how the AI MMI Trainer marks your practice responses.

How the /30 mark is built

The University of Auckland assesses MMI candidates across eight core attributes in total. These are the skills that matter most for studying and practising medicine:

1

Communication

How clearly and confidently you express your ideas, at a steady pace and tone.

2

Equity

Your awareness of fairness and social justice, and the factors that create health inequities.

3

Ethical and moral reasoning

How you weigh different sides of an issue and reach a fair, principled decision.

4

Critical thinking and quality of argument

How well you analyse information and build a logical, persuasive case.

5

Cultural safety and self-reflection

Respect for different cultures and insight into your own perspective and biases.

6

Collaboration in teamwork and leading in teams

How you work with others, build trust, and handle conflict.

7

Resilience and motivation

Your determination, ability to bounce back from setbacks, and care for your own wellbeing.

8

Mental flexibility and problem solving

Your ability to think on your feet and come up with creative, adaptable solutions.

Here's the key part: in a real MMI — and in the AI Trainer — each individual station only assesses three of these eight attributes, not all of them at once. The three chosen are the ones most relevant to that specific scenario.

For example:

A teamwork station might be marked on:

  • Collaboration in teamwork and leading in teams
  • Communication
  • Resilience and motivation

An ethics station might be marked on:

  • Ethical and moral reasoning
  • Critical thinking and quality of argument
  • Cultural safety and self-reflection

The AI looks at the content and intent of each question, picks the three attributes that best apply, then scores your response out of 10 on each one.

That gives your final station score of /30 (3 attributes × 10 marks).

Built and reviewed by high scorers

The scoring model isn't a static "set and forget" system. It's regularly examined, tested, and tweaked by the ICanMed MMI tutor team.

Every tutor on the team is a current UoA medical student who personally scored extremely well on their own UoA MMI — between 185 and 201 out of 210.

That means the AI's marking is continually checked against real interview experience. The team reviews how the AI scores responses, compares it to what actually worked (and didn't) in their own interviews, and calibrates it accordingly.

Think of it as ongoing quality control — the goal is feedback that feels realistic and fair, not a one-time formula left running on its own.

Important to know

This scoring system is indicative only.

The University of Auckland does not publicly release its actual marking scheme, exact attribute-weighting logic, or scoring methodology. Because of that, no external tool — including this one — can guarantee it replicates the real assessment process exactly.

The goal isn't to predict your real result. It's to give you a strong, realistic benchmark for practice and improvement — so you walk into your MMI better prepared, not with a false sense of certainty.

Why this matters

This mirrors how real UoA MMI assessors score: they focus on only the three relevant attributes per station, not a fixed checklist of every skill.

So when you get feedback on a station, it reflects genuine, realistic interview practice — not a generic score applied the same way to every question.

That's what makes the feedback useful: it tells you which specific skills to sharpen for the kind of station you just practised.