?
Социально-психологические факторы доверия искусственному интеллекту: состояние исследований
The article presents a theoretical review of literature from the last ten years devoted to the analysis of socio-psychological factors of trust in artificial intelligence. The widespread adoption of automated artificial intelligence systems, which is associated with expected economic growth, reduced resource costs, and the optimization of various work processes, often faces user distrust in new tools and a reluctance to transform traditional work processes. The combination of factors that reduce trust in artificial intelligence leads to low economic efficiency in the implementation of innovations, despite the wide range of technical possibilities. In addition to the importance of considering cognitive and affective factors of trust in artificial intelligence, socio-psychological aspects that determine the ethical and value acceptability of using automated AI systems are of particular significance. Theoretical analysis revealed that individualizing moral foundations are positively associated with trust in artificial intelligence in situations where the use of AI systems benefits society. Users who prioritize binding moral foundations demonstrate greater distrust of artificial intelligence and are less willing to delegate some tasks to automated assistants. The values of Openness to Change and Self-transcendence are more positively associated with trust in AI and digital innovation, excluding situations of high risk to life or social injustice. The values of Self-Enhancement are also positively associated with trust in AI tools, but mainly in situations where artificial intelligence simplifies the achievement of the user's goals or expands human capabilities to do so. There is a heterogeneous structure of relationships between Conservation values and trust in artificial intelligence, due to cultural characteristics. However, it should be noted that Conservation values are most often considered predictors of distrust in automated AI systems. The importance of considering value congruence between users and perceived AI profiles is discussed. For AI system developers, there is a need to pay special attention to the possibilities of adaptive personalized adjustment of the value profiles of generative models to users, which will lead to more effective human-machine interaction. The article highlights the areas of future research in this field as part of the development of a systemic model of trust in artificial intelligence.