Bridging technological innovation and socio-technical transitions: A generative AI model for predicting circular economy readiness

ABSTRACT The transition toward a circular economy (CE) is central to sustainable development and requires technological innovation to assess countries’ readiness for systemic change. However, existing CE assessment frameworks remain largely static, are prone to bias in qualitative evaluations, and are constrained by small sample sizes, limiting statistical efficiency, predictive robustness, and policy relevance. In this study, we develop a Circular Economy Readiness Prediction (CERP) Model that leverages Wasserstein Generative Adversarial Networks (WGANs) and multivariate statistical analysis to predict the transition capacity of EU-27 countries, addressing small sample size limitations and enabling robust testing and clearer insights into CE readiness. Using 12 Eurostat indicators across the CE domains, the model operationalizes the Circular Material Use Rate (CMUR) as a measure of circular performance. The proposed framework achieves strong predictive accuracy (MSE = 0.0367, MAE = 0.1507, RMSE = 0.1915) and reveals how trade integration, material dependency, and greenhouse-gas mitigation jointly shape CE outcomes. Using hierarchical clustering, we classified the EU-27 countries into four categories (Leaders, Fast Followers, Emerging Adopters, and Laggards) providing a differentiated basis for targeted policy and ESG investment strategies. Anchored in the Natural Resource-Based View, Institutional Readiness Theory, and Socio-Technical Transition Theory, the study advances a novel interface between AI-driven modeling and transition governance. The CERP model demonstrates how generative AI can enhance evidence-based decision-making for CE policy design, funding allocation, and monitoring under the European Green Deal, while also contributing to broader insights on data-driven sustainability governance beyond the EU.