Portrayals of artificial intelligence (AI)

: assessing patient reliance in high-risk medical decision-making

Abstract

As healthcare increasingly adopts artificial intelligence (AI) for decision support, the success of AI systems depends on user trust and willingness to rely on recommendations, especially in high-risk medical scenarios. This study used an experiment featuring different portrayals of AI4presented as code-based, morph, or human-like4to assess how these visual representations impact patients' trust in AI recommendations. The findings reveal that visual portrayals of AI has no significant effect on utilitarian motivation, interaction convenience, or task-technology fit. Instead, perceived competence emerges as the most influential factor in building trust, which in turn increases reliance on AI recommendations. The results highlight that users prioritize the functional competence and reliability of AI over aesthetic or anthropomorphic features, particularly when accuracy and trust are critical. The study offers important managerial insights for organizations integrating AI systems, emphasizing the need for transparency, accuracy, and reliability to foster trust. While aesthetic enhancements may attract initial engagement, they should not detract from delivering clear and reliable outputs. By focusing on competence and trustworthiness, organizations can enhance AI adoption, particularly in healthcare, where accurate decision support is crucial. This research underscores the importance of designing AI systems that not only meet technical standards but also maintain user trust, ensuring their effective use in decision-making processes.

Date of Award	23 Oct 2024
Original language	English
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Filipa de Almeida (Supervisor)