This study examines the psychological and technological factors of the acceptance and emotional well-being of 349 college students with autism who use generative artificial intelligence tools as sources of psychological and emotional support. Grounded on the model of Technology Acceptance Model (TAM), the research studies relationships among perceived empathy, perceived usefulness, ease of use, emotional trust, continuance intention, and overall well-being. Structural equation modelling found that the perceived empathy was a significant predictor of ease of use and usefulness, while emotional trust and continuance intention proved to be significant predictors of emotional well-being. To reinforce these results, predictive modelling performed with Orange Data Mining showed a good accordance between observed values and predictions, thus showing robustness of the model. Continuance intention was the best predictor for well- being, followed by emotional trust and ease of use. These findings highlight the importance of designing emotionally responsive, trustworthy, and accessible generative AI systems in order to support sustained engagement and psychological safety among neurodiverse students.
George et al. (Fri,) studied this question.