Algorithm to emotion: a three-stage model of user acceptance for AI-generated news platforms

With the rapid development of AI, Users’ acceptance of intelligent recommendation technology on news platforms directly affects the effectiveness and sustainability of using intelligent recommendation AI news. This study investigates user acceptance of AI-driven news recommendation platforms by applying the Artificial Intelligent Device Use Acceptance (AIDUA) model. Through a survey of 1,100 users, we examine how six AI-specific factors—social influence, perceived novelty, intelligence, accuracy, transparency, and fairness—shape performance expectancy and effort expectancy, ultimately influencing acceptance decisions. Results reveal that all six factors positively impact performance expectancy, while social influence, accuracy, transparency, and fairness reduce effort expectancy. Notably, perceived accuracy (β = 0.200) exerts the strongest effect, underscoring content quality as a critical driver of trust. Emotion mediates between cognitive evaluations and behavioral outcomes, with positive emotions enhancing acceptance and negative emotions amplifying resistance. The study advances theoretical understanding by extending the AIDUA model to AI journalism, highlighting the dual-path role of cognitive and affective evaluations.