How Organizations Choose Open-Source Generative AI Under Normative Uncertainty: The Moderating Role of Exploitative and Exploratory Behaviors
Suengjae Hong, Hakshun Ryee, Jin Xiaoyan, Daegyu Yang · Journal of theoretical and applied electronic commerce research · 2025
Open-source generative AI technologies offer transparent and customizable alternatives to proprietary AI systems, the concept of which closely aligns with the principles of open innovation. Organizations with strong open-source orientations may have greater absorptive capacity to adopt open-source generative AI technologies. However, adopting such technologies into the organizations is not always guaranteed because ethical, privacy, and regulatory concerns on open-source generative AI usage create normative uncertainty that can reduce organizations’ willingness to adopt the technology, particularly when it is used in customer-facing products or services rather than integrated into internal processes. This study draws on organizational learning theory and open innovation literature to examine how open-source orientation affects open-source generative AI adoption under normative uncertainty, and how this relationship depends on organizational exploiting and exploring behaviors. Using global survey data from the Linux Foundation, we test our hypotheses with ordered logistic regression and interaction effects. The results show that open-source oriented organizations are more likely to adopt open-source generative AI, but this effect weakens when normative uncertainty is high, especially in product-related use cases. These findings extend absorptive capacity theory by highlighting ethical ambiguity as a key moderating factor and provide practical insights into how organizations can responsibly approach open-source generative AI adoption.