AI Innovation Ecosystems and Sustainable Development: Cross-National Configurational Pathways to Sustainable Innovation Performance

Artificial intelligence (AI) is becoming increasingly important for sustainable development, yet the relationships between AI innovation ecosystems and sustainability-related performance remain insufficiently understood. Drawing on data from 121 countries, this study integrates Necessary Condition Analysis (NCA), fuzzy-set Qualitative Comparative Analysis (fsQCA), and Artificial Neural Networks (ANN) to examine the necessity, configurational associations, and predictive relevance of five ecosystem conditions: AI infrastructure, institutional environment, technological innovation, human capital, and innovation synergy. Sustainable innovation performance (SIP) is operationalized as a composite proxy incorporating economic, environmental, and social outcomes. The results indicate that no individual condition robustly satisfies the adopted necessity criteria across analytical approaches. Instead, four alternative configurations are associated with high SIP: resource–synergy-driven, institution-guided, talent-driven, and infrastructure-supported. The ANN results further reveal differences in out-of-sample predictive importance, with institutional environment and innovation synergy exhibiting nearly identical leading contributions. These findings extend the configurational understanding of AI innovation ecosystems and provide evidence for differentiated approaches to sustainable development across national contexts.