Exploration or exploitation? The double-edged effects of AI use modes on radical creativity: evidence from AI-enabled drug discovery

Artificial intelligence (AI) has been widely adopted in corporate innovation activities, yet prior research has not reached a consensus on whether it is more conducive to incremental or radical innovation. For radical innovation, which depends on highly novel ideas, realizing the value of AI depends not only on whether AI is adopted but, more importantly, on whether employees can use it creatively to transcend established knowledge boundaries. Drawing on environmental dynamism and the exploration-exploitation perspective, this study develops a theoretical model linking leaders' perceptions of environmental dynamism, employees' AI use modes, leader temporal myopia, and employee radical creativity in the context of AI-enabled drug discovery. We distinguish leaders' perceptions of environmental dynamism into perceived technological dynamism and perceived market dynamism, and employees' AI use into exploratory and exploitative AI use. The model is tested using partial least squares structural equation modeling and three-wave matched survey data from 368 leader-employee dyads. The results show that leader-perceived technological dynamism is positively associated with employee exploratory AI use, whereas leader-perceived market dynamism is positively associated with employee exploitative AI use. Exploratory AI use is positively associated with employee radical creativity, whereas exploitative AI use is negatively associated with it. Leader temporal myopia weakens the positive conditional indirect relationship of technological dynamism with radical creativity through exploratory AI use and strengthens the negative conditional indirect relationship of market dynamism with radical creativity through exploitative AI use. By adopting a knowledge-activity perspective, this study reveals the differentiated relationships between AI use modes and radical creativity and further explains how perceptions of the external environment, leaders' temporal orientation, and employees' AI use behaviors are differentially related. The findings provide a theoretical basis and practical implications for helping firms engaged in AI-enabled drug discovery guide employees toward effective AI use and enhance radical creativity.