Background: Automated closed captioning is increasingly used in medical education to improve accessibility and support learning for individuals of all abilities. However, the specialized vocabulary used in medical lectures is often difficult for closed captioning software to accurately transcribe, potentially affecting learner comprehension and creating significant barriers for deaf and hard-of-hearing students. SUNY Upstate Medical University utilizes YuJa Media to record and transcribe lectures taught to first- and second-year medical students. This study evaluates the accuracy of YuJa Media's automated captions and characterizes common transcription errors. Methods: A retrospective analysis was performed using 60 recorded lectures across 10 academic units in the preclinical medical curriculum. Automatically generated captions were compared with the spoken content and manually reviewed for accuracy. Captioning errors were identified and categorized into predefined groups, including medical terminology, drug names, anatomical structures, and common English words. Error frequency and error rates will be calculated overall and by category. Results: Preliminary findings suggest that medical terminology accounts for a disproportionately high frequency of automated captioning errors, with variability across lecture topics. Baseline accuracy with common English words appears high, raising hopes that adding a module of medical education terms could address the software issue. Discussion: This study will provide insight into the reliability of automated closed captioning in medical education and identify common sources of transcription errors. The findings will be used to develop a glossary of commonly mistranscribed terms that will be shared with YuJa Media to improve the accuracy of automated captioning. This study may inform strategies to enhance accessibility for learners and support more equitable delivery of medical education.