Existing research has primarily focused on the instrumental attributes of Generative AI (Gen AI), such as its usefulness and usage patterns, in influencing employee creativity. However, there is a lack of systematic exploration of Gen AI attitude-defined as the degree of support and willingness to adopt Gen AI-and their underlying mechanisms in shaping creativity. Grounded in role theory, this study uncovers the dual effects of Gen AI attitude on creativity. On the one hand, Gen AI attitude can reduce creativity through increased role ambiguity. On the other hand, it can enhance creativity by facilitating task crafting. Furthermore, this study introduces creative self-efficacy as a moderating factor, revealing that it mitigates the negative impact of Gen AI attitude on creativity via role ambiguity while strengthening its positive effect through task crafting. These hypotheses are tested through two complementary studies. Study 1 employs a multi-wave survey (N = 591) and structural equation modeling to validate the proposed relationships. Study 2 utilizes a large language modeldriven text analysis approach, using GPT to analyze 415 social media texts, further reinforcing the robustness of the findings. The results contribute to a deeper theoretical understanding of how Gen AI attitude shapes creativity, offering new insights into the intersection of technology-related attitudes and creativity.

Ma Lin, Ge Longqi, Li Yan(corresponding author), Wang Yetong. When employees meet gen AI: The double-edged effects of AI attitude on creativity [J]. Technovation, Apr 2026, DOI: 10.1016/j.technovation.2026.103495