Artificial intelligence (AI) tools are increasingly used in medical education, clinical care, research, and administrative work. The University of Colorado Anschutz School of Medicine (CUSOM) supports the responsible, transparent, and ethical use of AI to enhance learning and patient care while preserving academic integrity, clinical reasoning, and human judgment.
These guidelines summarize expectations for medical students in the Trek Curriculum and are aligned with the CUSOM Policy & Guidelines.
Core Expectations for Students
AI is assistive, not a replacement. AI tools may support learning, organization, and efficiency but must never replace independent clinical reasoning, required skill development, or professional judgment.
Human responsibility remains paramount. Students are fully accountable for all work they submit, all documentation they enter, and all clinical decisions they contribute to, regardless of AI involvement.
Transparency is required. When AI is used substantively in academic or clinical work where permitted, students must disclose the tool used and how it contributed to the final product, consistent with faculty instructions and course policies. Routine use of AI for grammar, spelling, or minor editing generally does not require disclosure unless specifically requested. Students and faculty are expected to be transparent about AI use when asked. Example disclosure statements are provided below:
CUSOM AI Disclosure Statement (example): ChatGPT‑4o was used to generate an initial outline and first draft of this paper. All content was subsequently revised, verified, and finalized by the student.”
When AI is used in any submitted academic work, students must disclose: the specific tool used (e.g., “ChatGPT-4o was used to draft the assessment and plan”) and how the tool was used (outline, first draft, editing, ambient listening, etc.)
Graduated use across training. Early learners are expected to develop foundational skills independently before using AI for augmentation.
Faculty and preceptor authority. Instructors and clinical supervisors may restrict, permit, or structure AI use in their courses or clinical settings. Course‑ or clerkship specific guidance always applies.
Permitted Educational Uses (When Allowed by Course/Clerkship)
- Study aids, knowledge review, or concept clarification
- Drafting outlines or improving clarity and organization of written work (with disclosure)
- Comparing independently written clinical documentation to AI generated examples for learning
- Brainstorming differential diagnoses or care considerations after independent reasoning
Prohibited Uses
- Submitting AI generated work without review, editing, and disclosure
- Using AI during exams or assessments unless explicitly authorized
- Entering patient, student, or confidential institutional information into unapproved AI tools
- Using AI before meeting required competency milestones where independent performance is expected
Data Privacy and Security
Protected data must never be entered into unapproved AI tools. This includes patient information (HIPAA), student educational records (FERPA), and confidential institutional data.
Only institutionally approved AI systems may be used for clinical or educational work involving sensitive information.
Students should assume that information entered into nonapproved AI tools may be retained or reused outside their control.
More Information
Questions? Contact the Office of Assessment, Evaluation, and Outcomes.
