Enterprise AI Analysis

Initial bivariate tests showed no significant differences across variables, challenging the five initial hypotheses (H1-H5). However, multivariate logistic regression identified social influence as the sole statistically significant predictor of adoption willingness (p = 0.027, Exp(B) = 3.172). Other factors like performance expectancy, effort expectancy, initial trust, elemental trust, perceived risk, and perceived threat did not show a direct statistical correlation with the intention to adopt AI in this context.

Structured interviews revealed a surprisingly low awareness of AI-based clinical decision support systems among medical personnel, with very limited actual deployment. Participants' knowledge often stemmed from external sources like the internet, not institutional training. This suggests that the quantitative 'null results' were largely due to a pre-adoption state of technological invisibility and institutional inertia rather than user indifference. Key concerns identified included whether AI systems truly 'worked' (considering patient psychology), data security, and operational reliability (e.g., network crashes). Occupational threat was consistently dismissed, with doctors affirming the importance of human evaluation complementing AI diagnosis.

The study highlights that the primary barrier to AI adoption is not individual practitioner reluctance, but the failure of hospital administrators and university medical teachers to acquire and formally introduce AI-based clinical decision support systems and provide adequate training. This institutional inertia means potential users often lack basic awareness of the technology's existence and practical application. The findings underscore the importance of management support and education in promoting AI take-up, suggesting that efforts should prioritize acquisition of appropriate AI software/hardware and comprehensive training for both new and current medical professionals.