Differences in the determinants of AI adoption across sectors and technological intensity

Purpose – As artificial intelligence (AI) becomes increasingly central to business innovation and competitiveness, recent years have witnessed a surge of empirical studies examining the factors that influence its adoption. Despite this growing interest, relatively little research has explored how the determinants of AI adoption differ across sectors and industry characteristics. This study addresses this gap by identifying key factors associated with AI adoption using the technology–organization–environment (TOE) framework. Design/methodology/approach – This study examines how these significant factors differ between the manufacturing and service sectors. It also investigates variations within the manufacturing sector by classifying firms into four groups based on technological intensity: high, medium-high, medium-low and low. The analysis is based on firm-level data from Korea. To address potential bias in rare event data, Firth's penalized logistic regression is applied to identify sector- and industry-specific drivers of AI adoption. Findings – The results show that big data utilization, organizational proactiveness in digital transformation and government technical support are statistically significant drivers of AI adoption across both manufacturing and service sectors. In addition, high-tech industries are more influenced by software-based infrastructure, while low-tech industries are more closely associated with hardware and equipment-related conditions. Originality/value – These findings offer practical implications for promoting AI adoption by highlighting the importance of policy approaches that address sector-specific differences in technological infrastructure, organizational capacity and external institutional support, as conceptualized within the TOE framework.