This study aimed to examine the effect of AI knowledge acquisition among developmental rehabilitation service providers on their acceptance of AI intervention tools and to explore the parallel mediating pathways of job stress and job efficacy in this ...
This study aimed to examine the effect of AI knowledge acquisition among developmental rehabilitation service providers on their acceptance of AI intervention tools and to explore the parallel mediating pathways of job stress and job efficacy in this relationship. An online self-report questionnaire was administered to developmental rehabilitation service providers via online community forums and KakaoTalk, and responses from 100 participants—whose work consisted of at least 60% developmental rehabilitation services—were used in the final analysis. Using JAMOVI 2.7, the collected data underwent reliability analysis, descriptive statistics, Pearson correlation analysis, Welch’s one-way ANOVA, multiple regression analysis, and parallel dual mediation analysis.
The main results are as follows. First, the acceptability of AI intervention tools was relatively high, with an average score of 4.29, while AI information acquisition scored an average of 3.47, indicating a level above average. Second, an analysis of differences in key variables based on general characteristics revealed a significant difference only in AI information acquisition according to monthly average income; no significant differences in AI intervention tool acceptability were found based on service area, highest level of education, monthly average income, or clinical experience. Third, a significant positive correlation (r = .581, p 〈 .001) was found between AI information acquisition and acceptance of AI intervention tools, and a significant positive correlation (r = .283, p 〈 .01) was also confirmed between job efficacy and acceptance of AI intervention tools. Fourth, multiple regression analysis revealed that only AI information acquisition had a significant positive effect on the acceptance of AI intervention tools (B = 0.431, β = .566, p 〈 .001). In contrast, job stress, job efficacy, and general characteristic variables did not have a significant effect on the acceptance of AI intervention tools. Fifth, while the direct effect of AI knowledge acquisition on the acceptance of AI intervention tools was statistically significant (Estimate = 0.424, p 〈 .001), the indirect effects via job stress and job efficacy were both insignificant, indicating that a parallel dual-mediation effect was not confirmed.
These results suggest that the acceptance of AI intervention tools among developmental rehabilitation service providers is more directly explained by their acquisition of information about AI and their actual experience using it than by job-related psychological variables such as job stress or job efficacy, or general characteristics such as service area or years of experience. Therefore, to increase the acceptance of AI intervention tools in developmental rehabilitation settings, it is necessary to prioritize expanding providers’ access to AI-related information and providing educational support that includes practical use cases and ethical usage guidelines.