Secure and accessible AI in nursing education: A modular agentic chatbot framework comparing ChatGPT-4o with an...

Area of Science:

• Artificial Intelligence in Nursing Education
• Digital Health Education
• Agentic AI Frameworks

Background:

• Widespread adoption of AI in nursing education is hindered by cost, data governance, and reliability concerns.
• Modular, retrieval-augmented generation (RAG)-based agentic AI frameworks offer potential solutions by enhancing accuracy, flexibility, and local control.
• These frameworks can improve learning experiences and address limitations of current AI integration in healthcare education.

Purpose of the Study:

• To develop and evaluate a modular agentic AI chatbot framework for chronic wound care education in nursing.
• To compare the performance of open-source and commercial AI configurations within this framework.
• To assess the framework's potential for enhancing nursing education and promoting equitable access to information.

Main Methods:

• A comparative experimental pre-implementation study was conducted using a validated 100-item chronic wound care question dataset.
• A modular, agentic chatbot framework utilizing RAG technology was developed and evaluated with both open-source and commercial AI models.
• Six experts rated chatbot responses on accuracy, relevance, clarity, and coverage using a 5-point Likert scale, with statistical analyses including linear mixed-effects modeling.

Main Results:

• Both open-source and commercial AI chatbot configurations provided clinically accurate, guideline-aligned responses for chronic wound care education.
• The commercial model scored higher overall (mean difference = +1.01), particularly in coverage (+0.39).
• The open-source configuration showed strong guideline adherence, with 56% of responses fully aligned with clinical recommendations, demonstrating its potential.

Conclusions:

• The open-source, on-premises AI agent performed comparably to commercial models like ChatGPT-4o.
• This open-source solution presents significant advantages in terms of cost, data security, and institutional autonomy.
• The framework's ability to support guideline-based instruction and enhance equitable access positions it as a valuable tool for nursing education.