Appendix N: Course Syllabus Template

This syllabus is a modular educational template, not an accredited course offering or a substitute for local curriculum, CME, privacy, and institutional review. Faculty may adapt the structure, readings, and assignments to their specific context and audience.

Learning Objectives

Faculty using this template should be able to:

  1. Assemble a clinically grounded AI curriculum from the handbook’s evidence, specialty, implementation, and policy chapters.
  2. Adapt the sequence, workload, assessment, and specialty coverage to the learners and local educational requirements.
  3. Separate documented cases from hypothetical exercises and require primary-source verification for factual claims.
  4. Evaluate learners on evidence appraisal, safety, workflow, and communication rather than enthusiasm for a particular technology.
  5. Review current accreditation, CME, privacy, and institutional requirements before offering a formal course.
Using This Template
  • 12-week format: Standard graduate semester structure
  • 6-week intensive format: Combine weeks as indicated for accelerated programs
  • Modular use: Individual weeks can be extracted for focused workshops
  • Specialty tracks: Choose 2–3 specialty chapters relevant to the program
  • Shorter formats: Four- and eight-session tracks, competency maps, journal club questions for every chapter, and assessment items are in Teaching with the Physician AI Handbook

What this provides:

  • Complete 12-week graduate seminar syllabus using the Physician AI Handbook as the primary text
  • Weekly structure: required readings, learning objectives, discussion questions
  • Assessment framework: evaluation critiques, implementation plans, policy briefs
  • A grading rubric for the written assignments and a policy on learners’ use of generative AI
  • Alignment with the AAMC principles for responsible AI use in medical education, and a plan for evaluating the course
  • 6-week intensive format for CME or accelerated programs
  • Adaptation notes for medical schools, residencies, CME, and informatics programs

Key pedagogical approach: Teach evidence boundaries, failure analysis, and clinical accountability before product enthusiasm.

  • Start with failures (MYCIN, Epic sepsis, Watson) before successes
  • Emphasize critical evaluation over AI enthusiasm
  • Include ethics, bias, and liability throughout
  • Balance specialty depth with implementation practicality
  • End with policy engagement and career pathways

Illustrative assessment weights: Participation 15%, evaluation critique 15%, midterm 20%, implementation plan 20%, and final brief 30%. Faculty should revise these weights to match the learners, credit requirements, and institutional policies.

Course Overview

Course Title: Clinical AI for Physicians: Evidence-Based Evaluation and Implementation

Level: Graduate seminar (MD, DO), residency elective, or CME program

Prerequisites: Medical degree or advanced clinical training. No prior AI knowledge required.

Format: One 2-hour seminar per week (24 contact hours over 12 weeks), with about 4 hours of reading and assignment work outside class. Set the credit value under the institution’s credit-hour policy, and award CME credit only through an accredited CME provider.

Primary Text: Tegomoh, B. (2025). The Physician AI Handbook: Peer-Reviewed Evidence for Every Specialty.

Learning Objectives:

By the end of this course, students will be able to:

  1. Explain core AI/ML concepts relevant to clinical practice
  2. Critically evaluate AI tools using peer-reviewed evidence
  3. Apply AI evaluation frameworks to specialty-specific applications
  4. Assess AI systems for bias, safety, and equity considerations
  5. Navigate regulatory and liability issues in clinical AI
  6. Design implementation strategies for AI tools in clinical workflow
  7. Communicate AI limitations and uncertainties to patients and colleagues

12-Week Syllabus

Week 1: History and Context of AI in Medicine

Theme: Why clinical AI keeps failing, and what those failures teach

Required Reading:

Learning Objectives:

  • Trace AI development in medicine from MYCIN to modern LLMs
  • Explain why selected MYCIN evaluations reported expert-comparable performance while the system did not achieve routine clinical deployment.
  • Identify patterns in clinical AI failures over 50 years

Discussion Questions:

  1. Selected MYCIN evaluations reported performance comparable to expert judgment on bounded cases. What prevented routine deployment, and what does the evidence not establish?
  2. What does Google Flu Trends show about the gap between validation and real-world performance?
  3. What is different about the current AI wave compared with previous cycles?

Week 2: AI Fundamentals for Clinicians

Theme: What physicians need to understand about AI

Required Reading:

Learning Objectives:

  • Distinguish supervised, unsupervised, and reinforcement learning
  • Explain why clinical AI faces unique data quality challenges
  • Assess the generalizability limitations of AI trained on specific populations

Discussion Questions:

  1. Why do AI systems trained at one hospital often fail at another?
  2. What clinical data quality issues are most likely to cause AI failures?
  3. How should clinicians interpret vendor accuracy claims?

Week 3: Evaluating AI Systems

Theme: How to assess clinical AI before adoption

Required Reading:

Learning Objectives:

  • Apply current TRIPOD+AI reporting and PROBAST+AI risk-of-bias frameworks to prediction-model studies
  • Identify red flags in vendor validation studies
  • Distinguish internal validation from external validation from prospective deployment

Discussion Questions:

  1. A vendor reports 95% sensitivity in its validation study. What questions should you ask?
  2. How do you evaluate AI when prospective randomized trials do not exist?
  3. What is the difference between a model working and a model helping patients?

Assignment: Evaluation critique (1,500 words): Critically evaluate a published clinical AI validation study

Week 4: Ethics, Bias, and Health Equity

Theme: How AI can harm patients and widen disparities

Required Reading:

Learning Objectives:

  • Analyze the Epic sepsis model’s external-validation limitations and identify what subgroup evidence would be needed for an equity claim
  • Identify sources of algorithmic bias in clinical AI
  • Apply ethical frameworks to AI deployment decisions

Discussion Questions:

  1. An external validation reports low sensitivity and positive predictive value for an implemented sepsis model. Which subgroup analyses, uncertainty estimates, and workflow outcomes are needed before drawing an equity conclusion?
  2. Is it ethical to deploy AI with known demographic performance gaps if it still improves average outcomes?
  3. How should informed consent work for AI-assisted clinical decisions?

Week 5: Privacy, Safety, and Liability

Theme: Legal and regulatory landscape for clinical AI

Required Reading:

Learning Objectives:

  • Explain FDA regulatory pathways for clinical AI (510(k), De Novo, PMA)
  • Analyze liability allocation when AI recommendations cause harm
  • Identify HIPAA implications of cloud-based AI tools

Discussion Questions:

  1. Which facts, duties, warnings, workflow decisions, contracts, and jurisdiction-specific rules could affect liability when an AI-supported decision contributes to harm?
  2. How should physicians document AI-assisted decisions?
  3. What is the regulatory status of ambient AI scribes? What are the risks?

Week 6: Specialty Deep Dive I (Choose Your Track)

Theme: AI applications in specific clinical domains

Faculty Note: Select one or two specialty chapters relevant to the program. Radiology is a useful route because FDA’s public AI-enabled device list is heavily represented by radiology and the chapter includes accuracy, workflow, randomized, observational, and guideline evidence.

Suggested Options:

Learning Objectives:

  • Evaluate specialty-specific AI tools against peer-reviewed evidence
  • Identify which AI applications have strong evidence and which rest on marketing claims
  • Apply general evaluation frameworks to specialty-specific contexts

Discussion Questions:

  1. Which AI applications in your specialty have the strongest evidence? Which are overstated?
  2. How would AI change clinical workflow in your specialty?
  3. What is the greatest risk of AI in your specialty?

Midterm: Take-home exam covering Weeks 1–6

Week 7: Specialty Deep Dive II

Theme: Specialties with more complex AI applications

Suggested Options:

Learning Objectives:

  • Analyze AI applications in time-sensitive or high-stakes settings
  • Evaluate the tension between AI speed and clinical judgment
  • Assess AI tools for conditions with subjective diagnostic criteria

Discussion Questions:

  1. Should AI triage recommendations override clinical intuition in the emergency department?
  2. How should predictive models handle life-or-death decisions in critical care?
  3. Can AI assess mental health without understanding human experience?

Week 8: LLMs in Clinical Practice

Theme: Large language models, chatbots, and documentation AI

Required Reading:

Learning Objectives:

  • Explain how LLMs work and their fundamental limitations
  • Evaluate LLM applications for clinical documentation
  • Identify hallucination risks in medical AI chatbots

Discussion Questions:

  1. When should physicians trust LLM outputs? When should they verify?
  2. How should patients be informed about AI involvement in documentation?
  3. What happens when an LLM introduces unsupported clinical information into a note, message, or decision process?

Week 9: Workflow Integration

Theme: Why technically good AI often fails in practice

Required Reading:

Learning Objectives:

  • Apply human factors principles to AI implementation
  • Identify workflow disruptions that cause AI abandonment
  • Design implementation strategies that increase adoption

Discussion Questions:

  1. Why do clinicians disable AI alerts? How can implementation improve this?
  2. What is the difference between a useful AI tool and a usable one?
  3. How should AI fit into existing EHR workflows?

Assignment: Implementation plan (2,000 words): Design an implementation strategy for an AI tool in your clinical setting

Week 10: AI Vendor Evaluation

Theme: Assessing vendor claims and negotiating contracts

Required Reading:

Learning Objectives:

  • Apply vendor evaluation frameworks to real AI products
  • Identify red flags in vendor marketing and contract terms
  • Negotiate meaningful validation requirements

Discussion Questions:

  1. A vendor shows impressive demo results. What validation should you require?
  2. Which data-use, performance, update, audit, incident, termination, and liability issues require institution-specific legal and procurement review?
  3. What should happen when AI performance degrades after deployment?

Week 11: The Future of Clinical AI

Theme: Emerging technologies and implications

Required Reading:

Learning Objectives:

  • Evaluate emerging AI technologies (multimodal AI, autonomous agents)
  • Analyze how AI might change physician roles over the next decade
  • Assess implications for medical education and training

Discussion Questions:

  1. Will AI make physicians more or less essential? In what ways?
  2. How should medical education change to prepare for AI-augmented practice?
  3. What clinical tasks should never be delegated to AI?

Week 12: Policy, Governance, and Career Implications

Theme: Shaping AI policy and professional development

Required Reading:

Learning Objectives:

  • Analyze current and proposed AI regulations affecting clinical practice
  • Identify career pathways in clinical AI
  • Evaluate global health implications of AI deployment decisions

Discussion Questions:

  1. What AI governance principles should guide clinical deployment?
  2. How can physicians influence AI policy and development?
  3. What responsibilities do wealthy health systems have regarding AI equity?

Final Assignment: Policy brief or implementation proposal (3,000 words)

Assessment Structure

The following weights are an internally coherent example, not a universal grading standard. Faculty should adapt them to course credit, learner level, accessibility, institutional policy, and the reliability of each assessment.

Component Weight Due
Class participation 15% Ongoing
Evaluation critique (Week 3) 15% Week 3
Midterm exam 20% Week 6
Implementation plan (Week 9) 20% Week 9
Final policy brief or proposal 30% Week 12

Grading Rubric

One rubric covers the three written assignments. Weight its rows to the assignment: the evaluation critique rests on evidence appraisal and source verification, the implementation plan on workflow and safety, and the final brief or proposal on all five.

Criterion Exceeds expectations Meets expectations Below expectations
Evidence appraisal Names the study design, population, comparator, and outcome; separates internal, external, and prospective validation; states what the evidence cannot show Names the design, the validation level, and the main limitations Treats reported accuracy as proof of patient benefit
Source verification Traces every factual claim to a primary source; labels vendor figures as vendor-reported; checks regulatory status against FDA records Cites primary sources for most claims and labels vendor figures Relies on press releases, vendor materials, or AI-generated summaries without verification
Patient safety and equity Names plausible failure modes, the patients most exposed to them, the subgroup evidence that exists, and how harm would be detected after deployment Identifies the main failure modes and equity concerns Omits safety or equity, or asserts either without evidence
Workflow and implementation Specifies who acts on the output and when, how overrides and documentation work, and how performance is monitored after go-live Describes the workflow and a monitoring plan Assumes adoption without describing who uses the tool or how
Communication Clear to the stated audience of clinicians, leaders, or patients; states uncertainty plainly; the recommendation follows from the evidence Clear, with minor gaps in stating limits Unclear or overstated, or the recommendation does not follow from the evidence

Participation is assessed on preparation and on contributions that use evidence, not on volume.

Learner Use of Generative AI

Adopt a written policy before the first session and align it with the institution’s academic-integrity and privacy rules. A default policy for this course:

  • Learners may use generative AI tools to search, brainstorm, and edit. They disclose each use at the end of the assignment: the tool, the task, and how the output was checked.
  • Learners verify every citation, statistic, and regulatory claim against its primary source. An unverified or fabricated citation is graded under source verification, whatever produced it.
  • Protected health information and other identifiable patient data stay out of any tool the institution has not approved for that data.
  • The instructor states which assessments exclude AI assistance and how that is checked. A take-home exam cannot rely on a prohibition alone.
  • Course requirements use only tools the institution provides to every learner.
  • Learners remain responsible for everything they submit.

The course itself teaches the skill this policy depends on. Week 8 covers how language models introduce unsupported clinical information, so one exercise can ask learners to verify and correct an AI-drafted summary of a published validation study.

Alignment With the AAMC Principles

The AAMC’s Principles for the Responsible Use of Artificial Intelligence in and for Medical Education (Version 2.0, July 2025) set out seven principles for medical schools (AAMC, 2025). The table shows where this course addresses each one. The school’s curriculum committee decides whether that coverage meets its own requirements.

AAMC principle Where the course addresses it
Maintain Human-Centered Focus Weeks 7 and 11: when clinical judgment should override a model, and which clinical tasks should not be delegated
Ensure Ethical and Transparent Use Weeks 4 and 8 and the learner AI-use policy: informed consent, disclosure of AI involvement in care, and disclosure of learners’ own AI use
Provide Equal Access to AI The learner AI-use policy: course requirements use only tools the institution provides to every learner
Foster Education, Training, and Continuing Professional Development Week 11 on medical education, the 6-week intensive format, and the CME adaptation
Develop Curricula Through Interdisciplinary Collaboration The co-teaching note in the adaptation notes below
Protect Data Privacy Week 5 on HIPAA and cloud-based AI tools, and the learner AI-use policy
Monitor and Evaluate Weeks 3, 9, and 10 on monitoring after deployment, and the course evaluation plan below

The AAMC published its national AI competencies for the full medical education continuum in October 2026, listed under AI in Medical Education readings. Teaching with the Physician AI Handbook maps the handbook’s chapters to those competencies and to the ACGME Common Program Requirements. AI in Medical Education covers the wider curriculum and accreditation context.

Evaluating the Course

Plan the evaluation before the course starts:

  • Learner feedback: a short mid-course check at Week 6 and an end-of-course survey on relevance, workload, and readings.
  • Learning: two comparable published validation studies, one appraised in Week 1 and one in Week 12, scored with the evidence-appraisal and source-verification rows of the rubric. The change between them measures what the course taught better than a final exam score alone.
  • Application: a follow-up three to six months later asking whether learners have evaluated, questioned, or declined an AI tool in clinical practice.
  • Currency: an annual review of the readings against the handbook’s Changelog, because regulatory status, product evidence, and reporting frameworks change within an academic year.

6-Week Intensive Format

For CME programs or accelerated courses, combine weeks as follows:

Intensive Week Standard Weeks Focus
1 1–2 History, fundamentals, data challenges
2 3–4 Evaluation frameworks, ethics, bias
3 5–6 Safety, liability, specialty deep dive I
4 7–8 Specialty deep dive II, LLMs, documentation
5 9–10 Workflow integration, vendor evaluation
6 11–12 Future technologies, policy, careers

Supplementary Reading Lists

AI Evaluation Frameworks

AI Bias and Equity

FDA Regulatory Framework

AI in Medical Education

Clinical AI Implementation

Adaptation Notes for Instructors

Co-teaching helps: informatics faculty for Weeks 2 and 8, ethics and legal faculty for Weeks 4 and 5, and health-system procurement or IT staff for Week 10.

For medical schools (MS3–MS4 elective): Emphasize evaluation frameworks (Week 3), ethics (Week 4), and practical LLM use (Week 8). Reduce policy content.

For residency programs: Focus on specialty-relevant chapters. Add hands-on workshops with the AI tools already in local clinical use.

For CME/professional development: Compress to the 6-week format. Emphasize practical evaluation skills and implementation strategies over theoretical content.

For informatics programs: Expand technical content in Weeks 2, 8, and 11. Add technical evaluation exercises.

License

This syllabus template is released under the same CC BY 4.0 license as the Physician AI Handbook. Faculty may adapt freely with attribution.