Generative Artificial Intelligence Use Policy for Medical Students in the Medical Curriculum

Jacobs School of Medicine and Biomedical Sciences, University at Buffalo 

Revised: March 31, 2026 

The University at Buffalo Jacobs School of Medicine and Biomedical Sciences recognizes that generative artificial intelligence (GenAI) is rapidly transforming the landscape of health care, medical education, and clinical decision-making. Therefore, the Jacobs School acknowledges the need to include AI in the learning experience of its students. Including AI in curriculum and practice will enhance the knowledge and skills necessary for students and graduates to become effective scientists and health care professionals in a continuously advancing technological society.

As we adopt emerging technology, it is important to provide guidelines for the responsible and effective use of these emerging technologies. Unsupervised reliance on AI may undermine the development of independent clinical reasoning, erode professional accountability, compromise patient safety, or violate standards of academic integrity.

This policy sets clear expectations for the responsible, ethical, and supervised use of AI by Jacobs School medical students. Its goal is to support innovation while safeguarding the essential goals of clinical education: to build robust medical knowledge, cultivate sound judgment, promote reflective practice, develop strong communication skills, and uphold humanistic care in a rapidly evolving digital landscape.

This policy applies to all medical students enrolled at the Jacobs School of Medicine and Biomedical Sciences, including use of GenAI in activities related to coursework, assessments, research, and clinical settings where students are engaged in direct patient care activities within hospitals, clinics, and other clinical learning environments.  This applies to all contexts in which students represent themselves as learners or professionals within the Jacobs School community, whether activities occur on-site, off-site, or via digital platforms. It covers the use of AI tools for clinical decision-making, communication, documentation, academic assignments, and scholarly output. Students must comply with this policy in addition to all applicable institutional policies, including those related to academic integrity, patient privacy, professionalism, and data security.

Many assignments in medical school are designed such that the process of completing the work is just as important as the final product. Assignments, presentations, case write-ups, and oral communications are designed to build critical competencies, including clinical reasoning, reflective practice, analytic thinking, and sound professional judgment.

The use of generative AI must not bypass or replace these developmental goals. Students are expected to engage fully and authentically with all assigned work, ensuring that anything submitted, whether academic, clinical, or scholarly, accurately represents their own effort and understanding.

When AI is used to support a task (such as editing, summarizing, or idea generation), that use must be transparent and consistent with the expectations specified in the course syllabus or the faculty mentor in the case of individual projects or other experiences.

Academic integrity extends beyond avoiding plagiarism, it reflects the habits of inquiry, self-awareness, and ethical responsibility that form the foundation of professional identity and safe, effective patient care. It also ensures the deliberate construction of a durable knowledge base, which remains essential to the development of independent clinical judgment and lifelong learning. 

Information shared with Gen AI is not private and could expose proprietary or sensitive information to unauthorized parties. Information provided in prompts to GenAI is often stored by the AI software to be used in future outputs and to train or re-train GenAI models. Consistent with copyright and intellectual property rights, students can only upload system or course session materials, presentation slides, learning guides, written feedback, de-identified patient data, practice exam questions or unpublished research data into AI systems that offer contractual data protection and have been approved by University at Buffalo Information Technology. UBIT ensures that all software and applications procured on behalf of the University have the appropriate privacy, security and accessibility protections in place. Currently, the only generative AI tools approved by UBIT is the version of Microsoft Copilot that requires a UBIT login and the Semantic Clinical Artificial Intelligence (SCAI) chatbot.

Students are free to use generative AI as a study aid and to complete specific assignments as specified by the instructor or in this policy. In doing so, they must carefully follow the instructional guidance of faculty in the use of these tools. AI-generated content can be inaccurate, misleading or entirely fabricated and may contain copyrighted material. Students must verify accuracy and validity of content through comparison with peer reviewed, academic literature and other trusted sources. Students are responsible for any inaccuracies, misinformation, biased, offensive, or otherwise unethical content resulting from the use of these tools. When used to complete assignments, AI outputs should be cited when using direct quotations and paraphrasing, as well as using the tool for tasks like editing, translating, idea generation, and data processing.

Use of AI-assisted clinical support tools that do not receive protected health information and whose function is to synthesize curated literature in response to clinical questions are permitted under this policy.  Students are also permitted to use clinical site-deployed tools (e.g., Epic-embedded AI, ambient documentation assistants) when approved by a supervising medical faculty. 

Examples of permitted uses:

• Studying and concept review: Using AI to explain or summarize lectures, clinical topics, disease frameworks, or practice questions, as a personal learning tool.
• Literature support: Summarizing academic papers, outlining research background sections, or brainstorming research questions, under the guidance of a research mentor or faculty advisor. 
• Presentation preparation: Generating slide outlines, visual aids, or draft text for teaching sessions, case presentations, or journal clubs, when explicitly permitted.
• Language editing and formatting: Using AI tools to check grammar, spelling, or structure of non-clinical documents such as reflection essays, research summaries, or application materials.
• Clinical knowledge-building: Consulting faculty-approved AI-based clinical decision support tools (e.g., Open Evidence) to deepen understanding of disease mechanisms, diagnostic reasoning, and treatment guidelines, with appropriate supervision and verification.
• Course- or clerkship-specific tasks: Completing assignments or assessments where the use of AI is explicitly allowed by the course or clerkship director, with parameters for acceptable use. 

Use of generative AI is strictly prohibited in any context where it could undermine clinical learning, compromise academic integrity, breach patient confidentiality, or introduce safety risks. Students must never rely on AI in place of their own clinical reasoning, documentation, or patient communication.

Examples of prohibited uses

• Submitting AI-generated work as original content in any graded, clinical, or evaluated assignment.
• Using AI to generate or assist in the creation of a history and physical (H&P), progress note, or other clinical documentation, regardless of whether patient data is included.
• Using AI as a substitute for one’s own knowledge acquisition, analysis, or self-reflection, particularly in any assignment that evaluates clinical reasoning, professionalism, or communication.
• Relying on AI tools in place of clinical decision-making, diagnostic formulation, or independent preparation for patient care activities.
• Entering protected health information (PHI) or other patient identifiers into AI platforms, which constitutes a violation of HIPAA and university policy.
• Using AI during clinical encounters, pre-rounds, oral presentations, or group discussions without the express permission and supervision of the clinical faculty.
• Uploading institutional content (e.g., lecture materials, clerkship assessments, OSCE cases, preceptor evaluations, or de-identified research data) into public or non-UBIT-approved platforms.
• Preparing research posters, abstracts, or presentations with AI-generated content unless explicitly approved by a faculty supervisor.
• Bypassing reflective or experiential learning tasks, such as patient write-ups, professional identity assignments, or narrative self-assessments, through the use of AI.

AI tools must never be used to input, analyze, or transmit identifiable patient information. Doing so constitutes a serious breach of HIPAA (Health Insurance Portability and Accountability Act) and may violate university, hospital, and legal standards. This includes typed, photographed, or dictated data containing names, medical record numbers, dates of service, or any other identifying elements.

HIPAA violations can result in disciplinary action, including failure of the clerkship, reporting to the Promotions Committee, and dismissal from the Jacobs School of Medicine and Biomedical Sciences.

If a student is unsure whether a tool or prompt risks violating HIPAA or institutional policy, they are expected to err on the side of caution and consult with their supervising faculty or clerkship director before proceeding.

Misuse of AI that undermines professionalism or safety may result in disciplinary action. Violations of this policy may be considered academic misconduct and will be reviewed under the University at Buffalo Academic Integrity Policy and the Jacobs School Code of Professional Conduct. Students are encouraged to seek clarification from course directors before using AI tools in a new or uncertain context.

Faculty are expected to clearly communicate course- or clerkship-specific expectations regarding the use of generative AI tools in both clinical and didactic environments. Course syllabi and clerkship orientation materials should include a reference to this policy and explicitly outline whether AI is permitted, restricted, or prohibited for specific assignments, activities, or assessments. The Office of Medical Education will provide suggested language to support consistent messaging across courses and rotations.

Faculty may not implement blanket bans on the use of generative AI without consultation with and approval from the appropriate Phase leadership.

In the clinical environment, faculty and residents serve as role models and guides in helping learners use AI responsibly. The faculty are responsible for ensuring that clinical site-deployed tools that they approve for student use are covered by a Business Associate Agreement between the clinical site and the vendor

Supervision should focus on four core responsibilities:

• Detecting when AI has been used in student work or clinical tasks,
• Evaluating whether that use is appropriate in the given context,
• Framing AI as a tool that can complement, but never replace, clinical reasoning, and
• Teaching students to integrate AI outputs with evidence, patient values, and sound judgment.

This framework supports the development of humanistic, thoughtful clinicians and aligns with published recommendations on supervision in the age of AI.¹

Given the evolving nature of AI, this policy will be reviewed frequently and updated as needed. The Office of Medical Education will track student comments and criticisms of this policy and inform the Curriculum Committee. Students are responsible for staying informed of any changes.

Please contact the course director or the Office of Medical Education for questions about acceptable AI use.