ClaraNP: Generative Interface for Nursing Education with Hallucination Mitigation

Gulustan Dogan, Henry Salkever

Track: Innovations in AI for Education (Day 1)

Paper Length: short-paper (2 pages + references)

Keywords: Large Language Models, False Information Mitigation, Nursing Education

Abstract: Education holds great potential for improvement from generative AI integration. Efforts to employ AI tools in curriculum creation and generation have been by tried by companies, but recent reports of hallucination rates among commonly used LLMs demonstrate a rate of erroneous output far too high for reliable deployment in medical education. We propose ClaraNP, a generative LLM fine-tuned from Falcon-7b open-source code for nursing education with internal hallucination and bias mitigation. We demonstrate our preliminary progress, articulate our plan for achieving further domain specificity, and explain our plans for in-class testing and integration.

Submission Number: 48