Can ChatGPT help you prepare for an MMI?

Yes, for some jobs. A general chatbot like ChatGPT can write original practice scenarios, ask follow-up questions and point out gaps in a transcript of your answer. It may get admissions facts wrong, and its feedback can't establish how a school's raters will score a response, especially against criteria the school hasn't published. Use it for practice material and critique, and check facts with official sources.

Where it helps

• Fresh practice scenarios. Ask for everyday, nonclinical dilemmas you haven't seen before.
• Follow-up questions. After you answer aloud, paste in a transcript and ask what a curious listener would ask next.
• Finding gaps. Ask whether your answer does what the prompt asks, and which reason is weakest.

Where it falls short

• School facts. It may describe formats, timings or requirements that are outdated or wrong. Acuity says each program sets up its video interviews differently. See Acuity's preparation guidance. Use the school's and the provider's own pages.
• Scoring. Each program chooses its own raters and rubric. See how video interviews are assessed. A chatbot's rating of your answer is its opinion, not a forecast.
• Delivery. Feedback on a typed transcript can't hear your pacing or pauses unless you describe them.
• Model answers. A polished answer written for you is easy to memorize and hard to adapt. It can make your real answers sound rehearsed.

Official sources checked October 18, 2026.

Prompts that keep it grounded

Write one short, original practice scenario for a medical school MMI. Make it nonclinical and low-stakes, with two reasonable options and no obvious right answer. Don't give me a sample answer.

Here is the prompt and a transcript of my spoken answer. Tell me in one sentence what I decided. Then name one part of the prompt I didn't address, and quote the sentence that came closest to addressing it. Don't rewrite my answer.

Ask me three follow-up questions about my answer, one at a time, and wait for my reply to each before asking the next.

These prompts keep the chatbot reacting to what you actually said, so it doesn't replace your answer with its own.

Keep it to practice

Use these exercises for independent practice with original or public prompts. Never use a chatbot to write or help with a response in an actual application or assessment. Each school sets its own rules. Queen's, for example, says using generative AI tools to obtain responses is prohibited at any assessment stage, and it describes disqualification for AI use in application and admission components. See the OMSAS Queen's guide. That rule doesn't say Queen's endorses or permits any particular practice method. If a school's policy leaves practice use unclear, ask the school.

Keep it private and fair

Don't paste questions from a real assessment you agreed to keep confidential. Leave out names and identifying details about other people in your stories. Check the tool's data settings before you share anything personal.

Speak first, then ask for critique

This exercise works from a transcript, while your MMI answer may be spoken. Record yourself, transcribe the recording, and give the transcript to the chatbot. Then record again with one change. Comparing your two recordings tells you more than the critique alone.

When to use what

Mentor feedback in Videoath has the same limit as any generated feedback. It helps you decide what to change, and it doesn't predict a result. For when a person may be worth paying, see the coaching comparison.

Choose a public MMI practice question, answer it aloud, and try one of the prompts above on your transcript.