# Determinants of acceptance and usage of generative AI among Chinese medical students: a UTAUT-based empirical investigation

## TL;DR

This study explores what influences Chinese medical students to use generative AI, finding that perceived usefulness and support are key factors.

## Contribution

The study extends the UTAUT framework to Chinese medical education and identifies age as a moderator in AI acceptance.

## Key findings

- Performance expectancy, facilitating conditions, and social influence strongly predict behavioral intention to use GenAI.
- Behavioral intention significantly mediates the relationship between UTAUT constructs and actual GenAI usage.
- Age moderates the relationship between effort expectancy and behavioral intention, as well as between behavioral intention and actual usage.

## Abstract

Generative artificial intelligence (GenAI) is rapidly transforming higher education, yet empirical evidence remains limited on the factors associated with its acceptance and usage among medical students, especially in non-Western, high-stakes educational contexts such as China. A clear and contextualized understanding of these mechanism is essential to effectively integrate GenAI into medical curricula and prepare future healthcare professionals for AI-augmented clinical practice. Grounded in the Unified Theory of Acceptance and Use of Technology (UTAUT) framework, this study systematically investigated the relationships between core UTAUT constructs, and Chinese medical students’ behavioral intention (BI) and actual usage (AU) of GenAI, testing direct, mediating, and exploratory moderated pathways.

A cross-sectional online survey was administered to students at a public medical university in China from October 2024 to January 2025, yielding 1781 valid responses. Validated scales were used to measure core UTAUT constructs: performance expectancy (PE), effort expectancy (EE), social influence (SI), facilitating conditions (FCs), BI, and AU. Partial Least Squares Structural Equation Modeling (PLS-SEM) was employed to test the hypothesized relationships.

The model demonstrated strong explanatory power, accounting for 67.6% of the variance in BI and 66.3% in AU. PE (β = 0.377, p < 0.001), FCs (β = 0.333, p < 0.001) and SI (β = 0.212, p < 0.001) were positively associated with BI. EE showed no significant direct association with BI (β = 0.038, p = 0.209) but had a weak yet significant direct association with AU (β = 0.057, p = 0.045). BI served as a significant mediator in the relationships between PE, SI, FCs, and AU (all p < 0.001) but failed to mediate the association between EE and AU (p = 0.219). Age was the only significant moderator for the path from EE to BI (β = 0.071, p = 0.043) and the path from BI to AU (β = 0.024, p = 0.022); gender, major, and academic level showed no moderating effects.

This study empirically validates and extends the UTAUT framework within Chinese medical education. Key findings underscore the important roles of PE, FCs and SI, reveal the context-dependent role of EE, and identify the moderating effect of age. Strategic interventions including demonstrating GenAI’s tangible utility, improving technical infrastructure, leveraging peer / faculty advocacy, and tailing strategies to age-related differences are recommended. These insights provide evidence-based guidance for educators, policymakers, and AI developers to support responsible integration of GenAI into medical education, ultimately preparing future healthcare professionals for an AI-driven healthcare ecosystem.