Machines in the Middle: Using Artificial Intelligence (AI) While Offering Help Affects Warmth, Felt Obligations, and Reciprocity

Abstract Theories of helping in the workplace are traditionally rooted in human interactions, often drawing from social exchange concepts. However, with the rise of artificial intelligence (AI), intelligent machines in work settings can now be leveraged in a growing number of areas. Given this change, we do not know how the use of AI-based tools will alter coworker perceptions and their subsequent responses to helping. Across two studies (and a replication), our results show that AI-assisted helping behavior, compared to unassisted help, is perceived as less warm, decreases felt obligations, and reduces one’s likelihood of reciprocating help. These results demonstrate how the integration of AI tools, which can enhance task efficiency and expand workers’ capabilities, might affect the social dynamics essential for effective cooperation. Theoretical and practical implications of these findings, as well as limitations, are discussed.