Workers are increasingly using AI to assist in the performance of work tasks. Although the self-directed benefits and drawbacks of AI usage have received scholarly attention, less is known about how co-workers react to a focal worker’s usage of AI, especially when such worker produces low quality outputs and makes mistakes. Interestingly, the impact of AI usage on the effect of poor performance could go either way: When using AI, one may be perceived either less responsible for the outputs, because part of the work is performed by another agent (i.e., the AI), or more responsible, because AI is known to hallucinate and AI users are expected to ensure that, when they use AI outputs, such outputs are truthful and correct. On a sample of 341 students, we tested our hypotheses in a co-working scenario where a colleague (i.e., peer) provided inputs necessary to one’s work performance. In support of the latter argument, we found that the negative effect of the peer’s poor work quality on peer evaluation by participants was amplified when the peer used AI. We identified perceived responsibility as a mediating mechanism of the moderation. Moreover, we found that more negative affect mediated the moderated effect of low work quality on peer evaluation. Thus, a peer’s poor performance has a more negative effect on peer evaluation through an affective mechanism when the peer uses AI, because the peer is deemed more responsible for the work quality. This finding holds important implications for work collaboration under AI usage.

Authors:
Federico Magni, Nanyang Technological University, Singapore
Ganqi Tang, Université de Fribourg, Switzerland

About the Presenter(s)
Federico Magni is an Assistant Professor of Organizational Behavior at Nanyang Technological University (Singapore). He studies creativity and AI in organizations.

Connect on Linkedin
https://www.linkedin.com/in/federicomagni90/

See this presentation on the full schedule – Wednesday Schedule

Bookmark