AI Education Atlas · November 13, 2026

Every framework teaches what AI is. Almost none proves the training worked.

I mapped ten healthcare AI education frameworks onto the six plain-language buckets of the University of Toronto T-CAIREM education mandate[1], plus four gap checks. New in this version: the AAMC competencies[9], published October 27, 2026, and the first framework built for emergency medicine[10].

What I found

Loading.

So what

The field agrees on what to teach. It has not agreed on how to prove it worked, who teaches the teachers, or how to teach AI agents that act on their own.

Now what

Measurement, governance and agents are where I focus my talks and advisory work. I add each new framework here as it publishes, so you can watch which gaps close.

Coverage at a glance

Basics and bedside use are well covered. Measurement, teacher training and agents are not.

Crosswalk

Each framework against each bucket

Tap a framework name to read it. Tap a row to see how each one covers it. Squares under each name show the strength of the method behind it.

Order:
Validated with data Structured expert input No method stated

The six buckets

What healthcare AI training should cover, in plain language

Codes come from T-CAIREM's 3x3 framework[1]. Use them as crosswalk keys, for example "what does T3.4 say".

Sources and credit

Each framework belongs to its authors

These are my plain-language summaries, not quotations. Read the original before citing any wording.

    Next sources to check and add

    • AI-RADS residency curriculum, cited by T-CAIREM.
    • The scoping review of AI curricula for students, residents and physicians, cited by T-CAIREM.
    • Hospital medicine and nursing AI curricula, once I confirm one exists and is published.
    • The peer-reviewed version of the emergency medicine framework, when it appears.
    Use it

    Ten questions I ask of every AI education framework

    1. Who is it for, and at what training level?
    2. What must a learner know, judge and do?
    3. How is it taught: lecture, cases, coding, simulation?
    4. How does it measure success?
    5. Who teaches the teachers?
    6. Does it cover consent, privacy, liability and when to stop the AI?
    7. Does it cover AI agents that act, not just suggest?
    8. Is it open access?
    9. How often is it updated, and by whom?
    10. What can your organization adopt, cite or challenge?

    Bring this to your hospital, school or conference

    I speak and advise on the gaps this atlas shows: proving AI training works, connecting it to governance, and preparing clinicians for AI agents. Know a framework I missed? Send it and I will rate it.

    Explore my talks · Contact me