PURPOSE The rapid integration of Large Language Models (LLMs) into health
professions education requires a strategic adoption strategy. This
longitudinal study establishes a baseline of Artificial Intelligence (AI)
engagement, confidence, and ethical awareness among pre-clinical medical and
dental students. Findings are currently guiding a university-wide workshop
series designed to empower students with skills for responsible AI use.
METHODS A cross-sectional survey was distributed across four A.T. Still
University schools (KCOM, SOMA, MOSDOH, ASDOH) as phase one of a longitudinal
study. This analysis focuses on the 143 pre-clinical student participants.
The survey assessed usage patterns, self-reported skill progression,
perceived efficiency impacts, and concerns regarding ethics and data
security.
RESULTS Data reveals a critical gap between adoption and proficiency. While
over 80% of the cohort utilizes AI , 68.7% reported their skills stagnated or
regressed over the past year. Proficiency struggles are evident, with nearly
half (49.1%) attributing AI failures to "poor prompt phrasing".
Consequently, 51.6% reported AI had a neutral or negative impact on test
preparation efficiency. Ethical uncertainty is widespread: 44.1% expressed
low confidence in data confidentiality , and 56% were neutral or unaware of
intellectual property rights. Accordingly, "Advanced prompt
engineering" (62%) was identified as the top training priority.
CONCLUSIONS Pre-clinical students are active but struggling AI users. The
dichotomy between high adoption and low skill progression—evidenced by
admitted prompting failures and stagnant study efficiency—indicates that
access alone does not guarantee competence. Furthermore, significant
confusion regarding data privacy poses institutional risks. These findings
underscore the urgent need for formal curricula focused on advanced prompting
and ethical application.