How MedRise uses AI
MedRise uses AI to help people practice clinical reasoning. It is for education only, it never sees patient information, and people make the decisions.
Our commitments
- Education only. MedRise tools are not medical advice and not clinical decision tools. They are never meant for decisions about a real patient.
- No patient information. Cases are synthetic. Do not type or say anything that could identify a patient.
- People decide. In program use, a faculty member reviews every learning plan, and AI output is never the basis for any academic or employment decision.
- Your data stays small. MedRise Viva does not receive or store audio and does not keep your answers. Progress stays in your browser. We do not train AI models on learner data without written consent.
- Answers can be wrong. AI can misjudge an answer. Every scorecard shows the expected answer, and you can report errors to [email protected].
Which services are involved
- Speech to text: your browser’s own speech recognition (in Chrome, run by Google). MedRise receives only the text.
- Answer checking: when enabled, an Anthropic Claude model compares your answer with the expected answer; otherwise a keyword check runs on MedRise’s server.
- Reading questions aloud: when enabled, ElevenLabs voices the case text; otherwise your browser’s built-in voice is used.
- MedRise AI Tutor: runs inside ChatGPT, a service of OpenAI, under OpenAI’s terms. MedRise is not affiliated with or endorsed by Anthropic, ElevenLabs, Google or OpenAI.
Principles we follow
We design with the AAMC Principles for the Responsible Use of Artificial Intelligence in and for Medical Education in mind: human-centered use, transparency, protecting data, equal access, and ongoing monitoring. This is our own commitment, not an endorsement by the AAMC.
Questions or a mistake to report? Email [email protected]. Last updated 2026-11-13.