BastionGPT: A tool providing clinical support for medical students and a new way of preparation for education and the medical field
Publication Date
Spring 2026
Presentation Length
15 minutes
College
College of Education
Department
Education, Department of
Student Level
Undergraduate
Faculty Mentor
Amanda Nelms
Presentation Type
Talk/Oral
Summary
Artificial intelligence is rapidly reshaping healthcare and education by streamlining tasks that normally require significant time and effort. BastionGT is a secure, healthcare focused AI platform designed to support documentation and provide intelligent assistance. Built on privacy‑preserving versions of leading models, it maintains HIPAA compliance and protects user data. Our project examines how AI can assist medical students and healthcare providers. While AI is helpful for studying, organizing information, and accessing reliable references, it also raises concerns about authenticity and responsible use. To explore this, we interviewed a doctor of education with clinical experience. She emphasized that AI should enhance, replace, the human skills essential to medicine, especially building meaningful patient connections. She also cautioned against using AI to generate academic work or clinical documentation without human oversight. She responded positively to BastionGT’s secure transcription features and its refusal to learn from user data, noting that reducing administrative burden could give clinicians more time with patients. However, she recommended introducing AI tools later in medical training so students first develop strong interpersonal and clinical reasoning skills. Overall, our goal is to show how thoughtfully integrated AI can support healthcare practice while preserving the human elements at its core.
Recommended Citation
balderas, rubi and Newborn, KaNiyah, "BastionGPT: A tool providing clinical support for medical students and a new way of preparation for education and the medical field" (2026). SPARK Symposium Presentations. 960.
https://repository.belmont.edu/spark_presentations/960
