Alleviating privacy concerns in intelligent systems: the joint effects of autonomy and anthropomorphism via user anxiety

Advances in sensor technology and artificial intelligence have accelerated the adoption of intelligent systems, which are capable of recognising users’ needs and providing functionalities beyond users’ capabilities. Despite the substantial benefits offered by these systems, users continue to express significant privacy concerns, which remain a major barrier to technology adoption. Although technological solutions can enhance privacy protection, users often rely on subjective heuristics when assessing privacy risks. Anthropomorphism, which fosters interpersonal relationships and social connectedness between users and intelligent systems, is widely used in smart systems to reduce psychological distance. However, prior research has reported inconsistent findings regarding the effects of anthropomorphism on users’ privacy concerns. Study 1 examines whether anthropomorphic roles influence users’ privacy concerns and how this influence interacts with varying levels of autonomy in smart home systems. Results show that anthropomorphic roles alleviate users’ privacy concerns in low-autonomy systems, but not in high-autonomy ones. Furthermore, users’ anxiety mediates the interaction effect between anthropomorphic roles and autonomy levels. Study 2 further compares the influences of different anthropomorphic roles (assistant vs. family). Results reveal that the family role, as a socially closer role, evokes greater anxiety and consequently leads to higher privacy concerns than the assistant role.