The Effect of AI-induced Outcome Favorability on Vigilance Performance—an Exploration Based on Attention Capacity Theory

Chen Zhao,Beijia Qu,Jinlai Zhou,Chen Lin
DOI: https://doi.org/10.1007/s12144-023-05286-6
IF: 2.8
2024-01-01
Current Psychology
Abstract:While AI (artificial intelligence) embedded in work scenarios provides a great interaction experience for employees, this favorable experience instead traps employees in a comfort zone, ending up with a reduced vigilance performance. Based on attention capacity theory, this study identifies the mechanism by which AI-induced outcome favorability affects vigilance performance. Pre-test and Study 1 found that AI-induced outcome favorability can adversely affect vigilance performance and that failure experience interventions can help mitigate this adverse effect. Study 2 and 3 explored the mechanisms in detail. Study 2 is a scenario-based experiment that found that a failure experience intervention attenuated the positive relationship between AI-induced outcome favorability and AI complacency. Study 3 further demonstrated that the failure experience intervention moderated the mediating role of AI complacency in the relationship between AI-induced outcome favorability and vigilance performance through a two-period follow-up survey. There was no crossover between subjects in these studies. In conclusion, we find that AI-induced outcome favorability negatively affects vigilance performance via AI complacency, and that appropriate intervention modalities (failure experience intervention) can effectively attenuate this negative effect. We respond to traditional concerns in automation about complacency in human-computer interaction and extend the study of complacency in human-machine interaction. Our findings guide the organizational management of AI-embedded work processes, inspiring companies to initiate training on how AI works and to establish a scientific investigation mechanism for potential disputes over rights and responsibilities.
What problem does this paper attempt to address?