← All topics

Topic

AI in Medicine.

AI in medicine isn't one thing — it's diagnostic decision support, ambient documentation, algorithmic triage, image classification, and generative tooling. Each has different evidence, different risks, and different failure modes. Dr. Runge writes and speaks on the gap between what's deployed and what's documented.

What's actually deployed

Diagnostic decision support is in many radiology and pathology workflows. Ambient documentation is rolling out across health systems. Algorithmic triage in EDs is mixed evidence. Generative drafts of patient notes and patient-facing communications are now standard. None of these is the AI that wins headlines.

The hard questions

  1. How was the model validated, and on whose data?
  2. What's the failure mode when the model gets it wrong?
  3. Who benefits financially from the model being adopted?
  4. Does the system disclose to patients when AI is involved?
  5. What does liability look like when the model and the clinician disagree?

Where to read more

Stay close to this thread.

Be first to know when The Longevity Switch Substack and the Michigan MOOC launch. One email, no spam.