GreyBrain Pulse · Published 9 Oct 2026
Rewrite it in Hindi. Then prove nothing changed.
The clinician · What’s possible · F3
A named list of exactly where your instruction lost its meaning — the step every translation tool skips.
Why this matters for your role
The clinician’s skill: ask which system, population, comparator and outcome were actually studied.
Your worked example
A fictional teaching exercise compares a task-specific imaging system evaluated in a screening programme with a general-purpose chatbot that has not been evaluated in that programme.
Put it into practice
Build a checklist covering intended use, population, comparator, outcomes and validation. Explain why evidence for the imaging system cannot establish the chatbot’s clinical performance. Do not diagnose or recommend treatment.
Clinician prompt
Rewrite it in Hindi. Then prove nothing changed.
A named list of exactly where your instruction lost its meaning — the step every translation tool skips.
Help me check whether a patient instruction survives translation. This is a practice exercise using fictional material, not patient care. Step 1. Rewrite the instruction below in simple Hindi, for someone who left school after class 8. Use the words people actually say, not textbook Hindi. Step 2. Translate your Hindi version back into English, literally, as if you had never seen my original. Step 3. Put my original and your back-translation side by side in a table and mark every place where the meaning changed, softened or disappeared — especially doses, timings, and the warning signs that should bring someone back. Step 4. Rate each change as harmless, confusing, or dangerous, and rewrite the Hindi to fix every dangerous one. Instruction: "Tab. Metformin 500 mg twice daily after food. Return immediately if you develop vomiting, drowsiness or reduced urine output. Review in OPD after two weeks with a fasting sugar." End with a short section titled "Your check" listing what I must verify myself before using any of this.
Synthetic practice example. Run it as-is and check the output against the supplied facts.
Your check
The comparison table is the model’s claim about its own output. Read the Hindi yourself, or give it to a colleague who speaks it every day. Any changed dose or timing is a failure, not a nuance.
Fictional material only. Never paste a real patient’s instruction, name or record into a public AI tool.
Sources inform this fieldnote; their publishers do not endorse GreyBrain. This exercise does not confer clinical competence or a certificate.