Medical services are getting automated and intelligent. An emerging medical service is the AI pharmacy intravenous admixture service (PIVAS) that prepares infusions through robots. However, patients may distrust these robots. Therefore, this study aims to investigate the psychological mechanism of patients’ trust in AI PIVAS.
We conducted one field study and four experimental studies to test our hypotheses. Study 1 and 2 investigated patients’ trust of AI PIVAS. Study 3 and 4 examined the effect of subjective understanding on trust in AI PIVAS. Study 5 examined the moderating effect of informed consent.
The results indicated that patients’ reluctance to trust AI PIVAS (Studies 1–2) stems from their lack of subjective understanding (Study 3). Particularly, patients have an illusion of understanding humans and difficulty in understanding AI (Study 4). In addition, informed consent emerges as a moderating factor, which improves patients’ subjective understanding of AI PIVAS, thereby increasing their trust (Study 5).
The study contributes to the literature on algorithm aversion and cognitive psychology by providing insights into the mechanisms and boundary conditions of trust in the context of AI PIVAS. Findings suggest that medical service providers should explain the criteria or process to improve patients’ subjective understanding of medical AI, thus increasing the trust in algorithm-based services.