Enhancing Learning in Graduate Nursing Education Through a Co-Designed AI Virtual Tutor: A Mixed-Methods Evaluation.

Area of Science:

• Educational Technology
• Artificial Intelligence in Education
• Nursing Education

Background:

• Large language models (LLMs) are increasingly adopted in higher education for self-directed learning.
• AI virtual tutors offer contextualized support in nursing education but require evidence on accuracy and alignment.
• Empirical evidence on the use and impact of AI virtual tutors in nursing education is limited.

Purpose of the Study:

• Evaluate the use and perceived impact of a co-designed AI virtual tutor in a graduate nursing course.
• Explore student utilization patterns, perceived benefits, and limitations of the AI tutor.
• Assess the AI tutor's influence on student learning and engagement.

Main Methods:

• Pilot study with a mixed-methods explanatory sequential design.
• AI tutor trained on course materials and integrated into the learning management system.
• Data collected via usage logs, Bloom's Taxonomy coding, surveys (AI self-efficacy, usability, learning impact), and interviews with students and teaching assistants.

Main Results:

• 651 interactions logged; usage peaked around assignments.
• Most interactions involved lower-order thinking, shifting to application/analysis later.
• Students reported high AI self-efficacy, moderate use, and mixed perceived usefulness; most found it enhanced learning and recommended future use.
• Students valued immediacy and accuracy; TAs noted efficiency gains.
• Challenges included usability, scope, privacy, and over-reliance.

Conclusions:

• Co-designed AI virtual tutors are feasible and valued for contextual relevance in graduate nursing education.
• Perceived usefulness of AI tutors can be variable.
• Findings support the responsible, pedagogically integrated implementation of AI tutors in graduate nursing programs.