Changing Student Perceptions and Adoption Readiness for Generative AI Chatbots in Sino-British STEM Education: A Two-Wave Repeated Cross-Sectional Study (2023–2025)
Year: 2026
Type: article
Study design: Observational Study
Abstract: Generative artificial intelligence (AI) chatbots are increasingly prominent in higher education, but temporal evidence on student perceptions remains limited in English-medium transnational STEM education. This repeated cross-sectional study surveyed students at a Sino-British engineering institution in Spring 2023 (Wave 1; N = 165) and Spring 2025 (Wave 2; N = 297), and found that familiarity (37.6% vs. 67.3%; OR = 3.43), willingness (55.8% vs. 77.8%; OR = 2.78), comfort (63.6% vs. 74.4%; OR = 1.66), and perceived overall learning enhancement (46.1% vs. 69.0%; OR = 2.61) were higher in 2025. In a respondent-level model combining Wave 1 with the Spring four-programme subset (analytic N = 426), survey wave remained associated with willingness after adjustment for programme and year (OR = 3.51, 95% CI 2.17–5.69, p < 0.001). In the Wave 2 Spring concern checklist (N = 297), accuracy (51.5%) and deduplicated over-reliance (50.2%) were the leading concerns; because the concern blocks differed in format and options, specific concerns cannot be shown to have increased or matured. The findings indicate higher population-level adoption readiness, not individual progression or routine use. “Concern maturation” remains a hypothesis requiring invariant instruments and panel designs.
Source: Education Sciences
Institutions: University of Leeds, Southwest Jiaotong University
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: gold
Estimated APC paid: $2,172
OpenAlex ID: W7204111192