PatientTranslator
A patient-facing app that takes an AI-drafted clinical note (from a portal, AI scribe, or telehealth visit) and rewrites it in plain language, with confidence flags on anything that looks suspicious — so patients can spot when the AI might be wrong.
Patients reading AI-generated visit notes, discharge summaries, or portal messages from their doctor
- Plain-language rewrite of any clinical note with a 'what this means for you' summary at the top
- 'Show your work' panel: highlights terms the model wasn't sure about, with a 'ask your doctor about this' button and a flag if a term looks hallucinated or inconsistent with the rest of the note
- Side-by-side view: original clinician note vs. AI-drafted addendum, with anything new clearly marked
- Family/caregiver mode: shares a single safe link so a parent or partner can read the note together without creating an account
The paper proves that small token changes in medical text can carry huge clinical consequences, and that current evals miss it. Patients are already receiving AI-drafted notes — giving them a 'second pair of eyes' is a near-term trust and safety win.
Provincial audit found 20 approved AI scribe vendors routinely hallucinate facts in patient-clinician conversations, and AI-related medical malpractice lawsuits are emerging — patients are demonstrably receiving error-prone AI notes.Sick and wrong: Ontario auditors find doctors' AI note takers routinely blow basic facts ↗AI-Related Medical Malpractice Lawsuits ↗
Searches for 'patient-facing AI note explainer app' and 'Epic MyChart AI generated note consumer tool' returned zero direct competitors — existing AI scribe players (ChartScribe, OmniMD) target clinicians, not patients reading notes.Best AI Medical Scribes in 2026: 15 Tools Reviewed by Expert ↗
Health-app revenue is concentrated in fitness/wellness per Business of Apps; clinical-literacy B2C is unproven and patients rarely pay for portal-adjacent tools — likely needs B2B2C distribution via health systems rather than direct subscription.Health App Revenue and Usage Statistics (2026) ↗
AI scribes are scaling into standard care, regulatory scrutiny and malpractice exposure are rising, and federal plain-language requirements have been on the books since 2010 — a durable trust-and-safety wedge.Plain Language Materials & Resources | Health Literacy | CDC ↗
Core LLM rewrite is easy, but handling PHI (HIPAA/BAA), sourcing notes without portal APIs, and reliably flagging 'suspicious' claims are non-trivial — feasibility is middling.