Mapping interdependencies among challenges to artificial intelligence adoption in drug discovery

Abstract Artificial intelligence (AI) offers significant potential to accelerate drug discovery, yet its adoption remains constrained by interdependent technical, organizational, and regulatory barriers. This study applies Fuzzy DEMATEL to analyze twelve challenges and uncover their causal structure using expert judgments expressed as triangular fuzzy numbers. Following normalization and CFCS defuzzification, Resistance to Change, Lack of Standardization, Limited Trust in AI Outputs, Talent Shortage, and Lack of Clinical Validation emerge as the primary causal drivers. These upstream barriers exert strong influence on downstream effects, including Regulatory Barriers, Biological Complexity, Integration with Legacy Systems, and Poor Data Quality. The findings indicate that several constraints commonly viewed as external are, in practice, manifestations of internal organizational and technical shortcomings. By providing a structured causal prioritization of barriers, this study offers actionable guidance for targeting high-leverage interventions in AI-enabled drug discovery. Findings reflect expert judgment within a North Cyprus context and should be interpreted as indicative organizational signals rather than universal prescriptions.