Algorithm Delegation

Abstract

Integrating insights from psychological distance and human-AI agency theories, this research explores embedded AI — algorithms physically integrated into tangible devices as a novel mechanism to help overcome resistance towards medical AI. Across four studies, including a large text-mining (n = 224,433 reviews) and three controlled experiments (n = 677 participants), we show that embedded AI reduces psychological distance between consumers and AI (Study 2B), which in turn enhances willingness to delegate medical decisions to algorithms (Studies 1-3). We further demonstrate that consumers expect better health outcomes when delegating decisions to embedded AI (Studies 2A, 2B, and 3). Moderation analysis further shows that these effects weaken when an analytical mode is primed using numerical accuracy information (Study 3). Our findings contribute to ongoing discussions on overcoming resistance to medical AI and offer implications for the design of AI-based technologies in medical services.