Faculty Acceptance of Generative AI in Higher Education: A Meta-Analysis of TAM and UTAUT Studies (2021-2025)

Abstract

This meta-analysis synthesizes evidence from 10 empirical studies (2021–2025) on faculty acceptance of generative AI in higher education. Following PRISMA 2020 procedures, 523 records were screened, and 10 studies met the inclusion criteria for quantitative synthesis. Using random-effects models (REML), we estimated pooled associations between perceived usefulness (PU), perceived ease of use (PEOU), and social influence (SI) with attitudes (ATT) and behavioral intention (BI). All included studies employed cross-sectional survey designs (total N = 3,006), noting that the cumulative N varies across pathways because not all studies reported all relationships. Pooled effects indicated the most significant associations for PU with ATT (r = 0.40) and BI (r = 0.26), with more minor pooled associations for PEOU and SI. Heterogeneity was substantial across pathways (I² = 71–94%). Publication bias diagnostics did not indicate systematic bias for most pathways; interpretation of SI → ATT remains cautious due to k = 3. Overall, the synthesis suggests that perceptions of usefulness and ease of use are correlates of faculty attitudes and intentions to adopt generative AI, while highlighting substantial contextual variability.