Understanding art and design teachers’ willingness to adopt artificial intelligence in teaching under resource constraints: a mixed-methods study on perceived usefulness, resource readiness, and creativity-related concerns

Introduction Artificial intelligence (AI) is increasingly used in higher education, but teachers’ willingness to adopt AI in creative disciplines may not be fully explained by the instrumental logic of the Technology Acceptance Model (TAM). In art and design education, judgments about AI may also involve concerns about originality, skill development, process-based learning, and pedagogical governance. This study examined art and design teachers’ intentions to adopt AI under resource-constrained conditions. Methods An explanatory sequential mixed-methods design was used. The quantitative phase included 101 valid questionnaires and tested relationships among perceived ease of use (PEOU), perceived usefulness (PU), resource readiness (RR), attitude (AT), and behavioral intention (BI) using structural equation modeling with 5,000 bootstrap replications. The qualitative phase involved interviews with eight art and design teachers to contextualize the quantitative findings. Results Model fit was mixed (CFI = 0.935, TLI = 0.920, SRMR = 0.059, RMSEA = 0.096). Bootstrap results supported PEOU → PU, PU → BI, AT → BI, and RR → PEOU, whereas PU → AT and RR → PU were not supported. PEOU → AT showed a significant negative relationship, contrary to the hypothesized direction. Although RR → AT was positive, its standardized coefficient exceeded 1 and discriminant validity between RR and AT was insufficient; therefore, this path was not substantively interpreted. Interviews indicated that concerns about originality, skill degradation, shortcuts, copyright, academic integrity, and governance may shape how teachers translate functional evaluations into overall attitudes. Discussion Several core TAM pathways were supported, but the model was not fully validated. The findings suggest that AI adoption in art and design teaching involves both instrumental judgments and discipline-specific professional, pedagogical, ethical, and resource considerations. Given limitations in model fit and discriminant validity, the results should be regarded as preliminary and context-specific evidence.