Location
University of Nebraska Medical Center
Event Date
3-31-2026
Abstract
Introduction While early research indicated students preferred human narration over artificial intelligence (AI) voices, more recent studies have yielded mixed results, with some finding no significant difference in learner preferences for module narration. The purpose of this study was to determine whether professional level health science students exhibit a significant preference between human and AI generated narration in educational e-modules. Methods Our study used a noninferiority design to assess whether AI voices are comparably effective to human voice. Study participants completed an e-module narrated by a human or AI voice then completed a 21-item survey of perception and cognitive load. Results Our project included 320 health professions students. The mean Facilitating Learning (FL) score using the AI voice was found to be significantly noninferior to the mean FL score for using the human voice (p = 0.020). Mental effort scores between the groups were not statistically significantly different, suggesting that cognitive load between narration styles was similar for the two groups. Conclusions The results show that while students slightly preferred human narration over AI, the AI narration remained within an acceptable margin of inferiority, indicating that the difference between narration styles is not large enough to suggest human narration should be used preferentially. Using AI text-to-speech narration could streamline educational content creation, reducing time and resource demands for educators.
Rights
The author holds the copyright to this work and any reuse or permissions must be obtained from the author directly.
Included in
Voices of Learning: Exploring the Impact of Human and Artificial Intelligence Voice in Health Professions Education
University of Nebraska Medical Center
Introduction While early research indicated students preferred human narration over artificial intelligence (AI) voices, more recent studies have yielded mixed results, with some finding no significant difference in learner preferences for module narration. The purpose of this study was to determine whether professional level health science students exhibit a significant preference between human and AI generated narration in educational e-modules. Methods Our study used a noninferiority design to assess whether AI voices are comparably effective to human voice. Study participants completed an e-module narrated by a human or AI voice then completed a 21-item survey of perception and cognitive load. Results Our project included 320 health professions students. The mean Facilitating Learning (FL) score using the AI voice was found to be significantly noninferior to the mean FL score for using the human voice (p = 0.020). Mental effort scores between the groups were not statistically significantly different, suggesting that cognitive load between narration styles was similar for the two groups. Conclusions The results show that while students slightly preferred human narration over AI, the AI narration remained within an acceptable margin of inferiority, indicating that the difference between narration styles is not large enough to suggest human narration should be used preferentially. Using AI text-to-speech narration could streamline educational content creation, reducing time and resource demands for educators.