A meta-analysis and systematic review of empirical studies on teachers’ use of artificial intelligence found that perceived usefulness had the strongest pooled association with behavioral intention, while perceived ease of use was also associated with teachers’ intentions to adopt AI. The systematic review covered 38 studies involving 15,919 teachers in 19 countries; the quantitative meta-analysis included 24 studies with extractable bivariate correlations and 11,768 teachers. The results describe associations rather than causal effects or teachers’ actual classroom use, and the relationships varied substantially across studies.
Reported in Artificial Intelligence Review , the study examined the Technology Acceptance Model, or TAM, which focuses on perceived usefulness, perceived ease of use and behavioral intention. These constructs had been studied widely in relation to technology adoption, but the researchers noted that AI differs from the static, deterministic technologies for which the model was originally developed. AI can be opaque, adaptive and capable of challenging teachers’ professional roles and identities.
The review therefore asked not only how strongly the central TAM variables were associated with adoption intentions, but also what contextual conditions appeared to accompany variation in those relationships. Following PRISMA guidelines, the researchers synthesized the full set of 38 studies and conducted a random-effects meta-analysis using the DerSimonian–Laird variance estimator. Across the 24 studies in the quantitative synthesis, the pooled correlation between perceived ease of use and perceived usefulness was r = 0.532, with a 95% confidence interval of 0.410 to 0.636.
Perceived ease of use was associated with behavioral intention at r = 0.536, with a 95% confidence interval of 0.435 to 0.623. The association between perceived usefulness and behavioral intention was strongest at r = 0.619, with a 95% confidence interval of 0.511 to 0.708. These pooled values characterize average relationships among teachers’ perceptions and intentions; they do not demonstrate that changing ease of use or usefulness would cause adoption.
Despite the moderate-to-strong average correlations, the study found extremely high statistical heterogeneity. The I 2 values ranged from 97.9% to 98.6%, and prediction intervals extended from negative values to values near unity. In practical terms, the average associations held directionally, but a given result from one context could not be expected to describe every educational setting.
The pooled estimates were robust to alternative variance estimators and to the inclusion of studies based on the Unified Theory of Acceptance and Use of Technology after a ruling that their relevant constructs were equivalent. Egger’s regression indicated funnel asymmetry for all three associations, but leave-one-out diagnostics attributed that pattern entirely to one large-sample, low-effect study. The authors cautioned that, under heterogeneity of this magnitude, the asymmetry could not be interpreted as evidence of publication bias.
The qualitative synthesis identified five dimensions associated with variation in teachers’ AI adoption judgments: AI literacy and trust as preconditions for usefulness judgments; a shift from functional acceptance toward identity-based acceptance influenced by pedagogical beliefs and institutional norms; sociocultural and institutional context; tool-specific adoption logics; and ethical-risk calibration as a values-based gate on adoption decisions. The authors argued that these findings motivate extended frameworks that account for AI’s sociotechnical complexity. The practical implication is bounded rather than prescriptive: teachers’ adoption intentions may involve more than perceived usefulness, ease of use and intention alone, but the review did not establish a causal model or demonstrate that any particular contextual factor reliably predicts adoption in a new setting.
Subject of Research: Teachers’ perceptions and contextual factors associated with intentions to adopt artificial intelligence in education
Article Title: Explaining teacher adoption of AI through the technology acceptance model: a meta-analysis and systematic review
Article References: Shenoy, P., & Saarela, M. (2026). Explaining teacher adoption of AI through the technology acceptance model: a meta-analysis and systematic review. Artificial Intelligence Review. https://doi.org/10.1007/s10462-026-11736-x
Image Credits: AI Generated
DOI: 10.1007/s10462-026-11736-x
Keywords: teacher AI adoption, artificial intelligence, education technology, Technology Acceptance Model, perceived usefulness, ease of use, behavioral intention, meta-analysis, systematic review, AI literacy, Explaining, teacher
News Source: Denise Maddox. (October 21, 2026). Research report: Explaining teacher adoption of AI through the technology acceptance model: a meta-analysis and systematic review. Scienmag.
