Abstract:

Objective To explore the usability of the technical carrier of an artificial intelligence-driven problem-based learning（AI-PBL） model. Methods Guided by situated cognition and cognitive scaffolding theories，this study constructed an AI-PBL model and developed a web-based system prototype serving as its technical carrier，which comprises three core modules：staged case presentation，structured recording，and conversational interaction. A descriptive cross-sectional design was adopted. From October to November 2025，the system was implemented in the circulatory system module of the Adult Nursing PBL for 2023 cohort undergraduate nursing students at a nursing school. Upon completion of learning with the assigned clinical case，students evaluated the system prototype using the Chinese version of the System Usability Scale（SUS） and provided usage feedback by answering four open-ended questions. Results A total of 43 students were included. The SUS total score was 73.37±9.30. Students rated items related to function integration and intention to use the system relatively higher，and open-ended feedback mainly focused on optimizing scenario progression explainability，interaction stability，and instructional guidance support. Conclusion The system prototype of the AI-PBL model exhibits good usability and learner acceptability in Adult Nursing PBL. Further studies using controlled or longitudinal designs are needed to examine its effects on clinical reasoning and self-directed learning outcomes，and to further optimize system data security and formative evaluation mechanisms.

Key words: Generative Artificial Intelligence, Problem-Based Learning, Clinical Decision Making, Education, Nursing, System Usability