Learning Path 2

Use AI in Medicine

Move from isolated prompts to observable clinical workflows with deliberate human checkpoints.

5 courses 24 lessons 238 minutes
Who this is for
Physicians who already use ChatGPT or another general-purpose AI tool.
What you will leave with
Design an AI-assisted workflow that protects clinical judgment, data quality, and accountability.

A prompt completes a task. Clinical work is a sequence of tasks, decisions, handoffs, and checks.

This path examines that difference. You will learn to map the workflow, define the human checkpoint, build the log before the intelligence, and choose infrastructure that matches the sensitivity of the work.

Course Sequence

Begin in order.

The publication date does not determine the sequence. The learning dependency does.

Course 1

Clinical AI Workflows

Design AI-assisted clinical work as an observable sequence rather than a collection of prompts.

intermediate 46 min 5 lessons

What you will learn

  • Distinguish an AI task from an end-to-end workflow.
  • Place human checkpoints where clinical judgment belongs.
  • Define the logs required before automated interpretation.
  • Choose an infrastructure boundary that matches the data and use case.

Prerequisites

  • Routine use of at least one general-purpose AI assistant.
  1. 01 From Tasks to Workflows Identify the coordination work that remains after an AI model completes an individual task. 11 min
  2. 02 Protect Clinical Sequence Design documentation automation around the order of clinical work and the judgment it must preserve. 8 min
  3. 03 Logs Before Intelligence Establish clean, structured, timestamped ground truth before adding interpretation. 8 min
  4. 04 Choose the Infrastructure Boundary Compare consumer AI with locally controlled models through privacy, governance, cost, and operational requirements. 12 min
  5. 05 Delivery Date AI and the Meaning of a Due Date Apply clinical workflow principles to an image-derived prediction while preserving its meaning, limitations, and human review. 7 min

Course 2

Ambient Clinical Documentation

Build a vendor-neutral ambient documentation workflow, from pre-encounter preparation through post-encounter verification, grounded in thirty years of clinical documentation history.

intermediate 67 min 10 lessons

What you will learn

  • Explain how clinical documentation evolved from paper charts to ambient AI.
  • Build a pre-encounter, encounter, and post-encounter ambient workflow.
  • Review AI-generated documentation critically before it enters the medical record.
  • Evaluate ambient platforms using durable, vendor-neutral criteria.

Prerequisites

  • Routine use of at least one general-purpose AI assistant.
  1. 01 Why Ambient, Why Now? Understand why ambient clinical documentation matters beyond faster note creation. 6 min
  2. 02 From Paper Charts to Ambient AI Place ambient AI within the longer history of clinical documentation, from 1992 paper charts to the present. 8 min
  3. 03 What Ambient Clinical Documentation Is Understand the ambient documentation pipeline without tying that understanding to a particular product. 6 min
  4. 04 The Three-Phase Ambient Workflow Learn a repeatable, vendor-neutral pre-encounter, encounter, and post-encounter workflow. 8 min
  5. 05 The Patient Conversation Use ambient documentation to improve, rather than degrade, the clinical conversation. 6 min
  6. 06 Review, Verify, and Sign Develop a disciplined post-encounter review habit built around clinical consequence. 6 min
  7. 07 Teach the Tech, Not the Vendor Learn to evaluate ambient technology without becoming locked into a single product. 7 min
  8. 08 Optimizing Your Ambient System Move from simply using ambient documentation to deliberately designing its output. 6 min
  9. 09 Build Your Own Practice Lab Experience the ambient documentation pipeline without purchasing a clinical platform. 7 min
  10. 10 The One-Week Ambient Challenge Evaluate whether an approved ambient workflow improves your work in real practice. 7 min

Course 3

Clinical AI Literacy: From Prompt to Governed Workflow

Learn to specify, inspect, compare, test, and govern the systems that place large language models inside clinical workflows.

intermediate 65 min 5 lessons

What you will learn

  • Write a clinical prompt as a bounded specification with explicit evidence and stopping conditions.
  • Inspect the context, tools, memory, permissions, and logs surrounding an AI model.
  • Compare agent platforms through durable architecture rather than temporary model rankings.
  • Analyze clinical AI failures across unsupported additions, factual alterations, and omissions.
  • Design a minimum viable clinical harness with meaningful physician supervision.

Prerequisites

  • Completion of Clinical AI Workflows or the ability to map an AI task from source data to human checkpoint.
  1. 01 The Prompt Is a Clinical Specification Define the source, context, objective, prohibitions, and evaluation that govern a clinical AI task. 12 min
  2. 02 Learn the Harness, Not Just the Model Inspect the context, tools, memory, permissions, logs, and human checkpoints that turn a model into a working system. 11 min
  3. 03 Same Model, Different System Compare agent platforms through context, tools, memory, permissions, observability, execution boundaries, and failure behavior. 13 min
  4. 04 The Anatomy of a Medical AI Hallucination Trace unsupported additions, factual alterations, and omissions from source input through clinical consequence. 14 min
  5. 05 Build the Minimum Viable Clinical Harness Design a bounded clinical AI workflow that can retrieve, generate, verify, challenge, stop, and preserve a useful record. 15 min

Course 4

Agents 101 for Physician Developers

Move from AI conversations to bounded, observable, and verifiable agentic work without surrendering clinical judgment.

beginner 35 min 3 lessons

What you will learn

  • Distinguish chatbots, AI-assisted workflows, and agents through the agent loop.
  • Identify the model, harness, tools, context, permissions, evaluations, and human checkpoints in an agentic system.
  • Delegate a small development assignment with explicit boundaries and acceptance criteria.
  • Supervise an agent from clinical workflow observation through a tested software prototype.
  • Match agent autonomy to privacy, verification, governance, and clinical risk.

Prerequisites

  • Routine use of at least one general-purpose AI assistant; no coding experience is required for the first lesson.
  1. 01 From Chatbot to Agent Recognize the agent loop, choose an appropriate level of autonomy, and identify the human checkpoints a physician must retain. 10 min
  2. 02 The Harness Is the Workplace Compare agent environments by execution location, context, tools, permissions, and review surfaces instead of product rankings. 12 min
  3. 03 Your First Agentic Workflow Delegate one small software assignment with explicit scope, approval points, tests, and evidence for review. 13 min

Course 5

The Middle Path: Evidence, Systems, and Physician Judgment

A multipart assessment of AI claims and clinical workflows through evidence, system controls, and accountability.

intermediate 25 min 1 lessons

What you will learn

  • Separate observed AI failures and capabilities from extrapolations and forecasts.
  • Apply the three-error check to a synthetic clinical draft.
  • Assess a workflow through evidence, permissions, verification, and accountability.
  • Write a reasoned decision with explicit reconsideration conditions.

Prerequisites

  • Familiarity with clinical documentation and routine use of an AI assistant; no coding required.
  1. 01 Doomers, Dismissers, and the Physician Between Them Use evidence, system design, and accountability to assess AI claims and a fictional discharge workflow. 25 min