Benevolence in Human-Robot Interaction: Conceptual Fragmentation and Charting a Path Toward Integration

Robots are increasingly deployed as collaborative partners, so judgments about their benevolence matter for trust and acceptance. Yet benevolence remains conceptually fragmented. Different frameworks define it differently, operationalize it with non-comparable instruments, and rarely acknowledge that mismatch. This paper compares selected frameworks to diagnose the sources and consequences of that fragmentation. Across them, the same basic question recurs: does this agent intend good toward me? The frameworks diverge, however, in level of analysis, causal mechanism, and measurement. The result is construct confusion that limits cumulative knowledge building. In response, we outline a research agenda centered on measure-and-item auditing, conceptual integration, and a more parsimonious operationalization of benevolence.