<![CDATA[This study aims to examine the effect of generative artificial intelligence (AI) on business decision quality among MSMEs in East Java, with cognitive risk as a mediating variable. This study is motivated by the increasing adoption of generative AI in supporting business activities, including information retrieval, strategic alternative development, and decision recommendations. Although generative AI can improve information accessibility and decision-making efficiency, its use without critical evaluation may generate cognitive risks due to excessive reliance on AI-generated outputs. This study employs an explanatory quantitative approach involving 100 MSME actors who use generative AI for business purposes in East Java. Data were collected using a five-point Likert-scale questionnaire and analyzed using Partial Least Squares-Structural Equation Modeling (PLS-SEM) with SmartPLLS 4. The results indicate that generative AI has a positive and significant effect on business decision quality (?=0.810; t=13.737; p<0.001) and a negative and significant effect on cognitive risk (?=-0.751; t=12.961; p<0.001). Furthermore, cognitive risk has a negative and significant effect on business decision quality (?=-0.140; t=2.236; p=0.025) and mediates the relationship between generative AI and business decision quality (?=0.105; t=2.086; p=0.037). The R² values indicate that cognitive risk is explained by 56.4%, while business decision quality is explained by 84.7%. This study contributes theoretically to the human-AI collaboration literature by revealing the cognitive mechanism underlying the relationship between generative AI utilization and business decision quality. Unlike previous studies that predominantly emphasize the direct benefits of AI for efficiency and productivity, this study demonstrates that AI-supported decision quality is also influenced by users’ ability to evaluate and manage cognitive risks. Practically, this study highlights that MSMEs should utilize generative AI as a decision-support tool through information verification processes and the reinforcement of human judgment to ensure effective and responsible technology adoption.]]>
