An AI-Facilitated Virtual Teaching Assistant for Undergraduate Nursing Honors Research: An Embedded Mixed-Methods Evaluation

BACKGROUND: Artificial intelligence (AI) is increasingly integrated into higher education. However, evidence on theory-informed AI interventions supporting nursing research training remains limited. PURPOSE: To evaluate an AI-facilitated teaching assistant (INSPIRE-AI) on final-year undergraduate nursing honors students' research self-efficacy, motivation, and research interest, and explore students' experiences of using INSPIRE-AI. METHODS: An embedded mixed-methods study comprising a one-group quasi-experimental pretest/posttest design with postintervention semistructured qualitative interviews. Nursing students received access to INSPIRE-AI throughout the honors year. Quantitative survey data (N = 146) were analyzed using paired t-tests and repeated-measures general linear models. RESULTS: Research self-efficacy improved significantly (P < .001). Students reported that INSPIRE-AI supported structured thinking and reduced uncertainty, though engagement varied due to trust concerns, perceived surveillance, and preference for familiar AI tools. CONCLUSIONS: Together, these findings suggest that INSPIRE-AI has the potential to support research self-efficacy through structured scaffolding; however, this interpretation should be considered alongside the broader educational support that students received throughout the honors program.