Trustworthy AI-Augmented Objective Structured Clinical Examinations in Nursing Education: Taiwan-Japan Viewpoint on 5...

Area of Science:

• Nursing Education
• Artificial Intelligence in Healthcare
• Assessment and Evaluation

Background:

• Generative artificial intelligence (AI) presents challenges in high-stakes assessments regarding governance, validity, and faculty preparedness.
• Integrating AI into nursing objective structured clinical examinations (OSCEs) requires a structured approach to ensure reliable and valid outcomes.

Purpose of the Study:

• To outline a pragmatic and transferable approach for integrating generative AI into nursing OSCEs.
• To establish a governance framework for AI-augmented OSCEs, ensuring human oversight, transparency, ethics, safety, and traceability.
• To propose a joint roadmap and shared registry for benchmarking AI-augmented OSCEs.

Main Methods:

• A 5-AI-role model (learning assistant, AI-augmented standardized patient, assessment assistant, case generator, learning analyst) mapped across pre-OSCE, peri-OSCE, and post-OSCE workflows.
• Leveraging Taiwan's agile development and staged pilots with A/B comparisons and explainability-by-design logging.
• Utilizing Japan's policy scaffolding, including national AI guidance and revised nursing curriculum with assessment blueprints.

Main Results:

• Distilled 4 cross-cutting governance pillars: human oversight, learning process transparency, ethics and safety, and traceability.
• Developed implementable techniques such as machine-readable rubrics, standardized patient persona cards, bias monitoring, and faculty development.
• Proposed a phase-gated governance blueprint for AI-augmented OSCEs, aligning with international AI principles.

Conclusions:

• A Taiwan-Japan collaboration offers a pragmatic, transferable model for trustworthy AI-augmented nursing OSCEs.
• The proposed governance blueprint addresses key challenges in AI integration, promoting reliability, validity, equity, and manageable workload.
• This Asia-Pacific reference model can guide institutions globally in adopting AI in nursing education assessments.