Selected work
My research connects questions that are often separated: how clinicians use new tools, how expertise develops, and how organizations can turn evidence into responsible practice.
01
AI tools become consequential when they enter real clinical environments. I study how clinicians adopt these tools, how they redistribute documentation and cognitive work, and what evidence leaders need to judge their effects.
Carl Preiksaitis and colleagues described real-world adoption of an ambient AI scribe and compared documentation time and note characteristics using electronic health record audit logs.
View publication02
Medical education has to prepare clinicians to work with AI while keeping the reasoning that supervisors need to see visible. This work moves from mapping the field to practical guidance for programs and teachers.
Carl Preiksaitis and Christian Rose synthesized the early literature and identified critical evaluation, assessment, and human–AI interaction as priorities for study.
View publicationGuidance for supervisors supporting appropriate AI use while preserving visible reasoning and learning.
View publicationA framework for matching reliance on AI to the stakes of an educational decision and the evidence available.
View publication03
Clinical data can make parts of work and learning visible at a scale that observation alone cannot. I direct SIERRA, the Stanford Integrated Emergency Research Repository and Archive, to support collaborative research in emergency medicine. My work in educational measurement asks what clinical documentation can tell us about the development of expertise.
Visit SIERRACarl Preiksaitis and colleagues used clinical documentation to examine the breadth and timing of resident clinical exposure across training.
View publicationCarl Preiksaitis and colleagues developed an instrument for evaluating the quality of emergency medicine learning resources.
View publication04
I work with clinical and educational audiences who want a grounded way to think about AI, judgment, and implementation.
Inquire about a talk
How educators can preserve visible reasoning, meaningful supervision, and appropriate independence.
What adoption, audit-log, and workflow data can reveal—and what they cannot.
How implementation science connects evidence, human experience, and the daily work of clinical teams.
Selected past talks