GREYBRAIN FIELDNOTES

Your many roles.
One method underneath.

One stream shows you something you can make this week. The other looks at what is being claimed, and what the evidence actually supports. Both end at a skill you can practise.

WHAT’S POSSIBLE · A useful workflowWHAT’S ADVANCING · Evidence in context

THE EDITORIAL LIBRARY · RECENT FIELDNOTES

What changed. What you can do about it.

Nothing matches these filters yet.

Choose another role or stream, or work through one of the standing exercises below.

THE EVERGREEN PRACTICE LIBRARY

Choose a role. Practise a skill.

Eight exercises that are always here, whichever hat you are wearing today.

EVERGREEN PRACTICE · THE CLINICIAN

Know what the AI evidence actually supports

Good questions matter. Practise using GenAI to organise uncertainty and examine a line of reasoning with greater care.

The moment you recognise

A teaching case with incomplete information and more than one explanation.

Try it with this synthetic material

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.

Make something useful

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.

Your check: Separate supplied facts from suggestions. Open the source, identify what is missing and record what you changed.

TWO MINUTES · NOTHING TO SIGN UP FOR

Your first win
starts here.

An AI draft can overclaim confidently, in fluent English, with no warning that it has done so. This is the check that keeps the doctor in the loop.

No sign-up. You keep the template.

THE JOURNAL CLUB LABGuided demo · fictional case

01Read the source
02Spot the leap
03Keep the method

First, what does the source actually say?

FICTIONAL STUDY BRIEF

48 residents attended a journal-club workshop. Their average appraisal quiz score rose from 6/10 before the workshop to 8/10 immediately afterwards.

No comparison group. No later follow-up. Patient outcomes were not measured.

Now imagine handing this to an AI and asking for a summary for tomorrow’s journal club. Your job is what comes next: checking what it hands back.

This example is prewritten. No live AI request or real study is involved.

Turn a useful habit
into something you can show.

Self-paced foundations. Private capstone work. An exam and assessed progression towards the IIHMR Bangalore certificate programme. Eligibility and programme terms apply.

Explore the learning and certificate pathway ↗