Governing AI sustainability: risks, current responses, and pathways for improved governance

Artificial intelligence development presents significant sustainability challenges across environmental, social, and economic dimensions, demanding more coordinated and systematic governance approaches. However, existing governance responses remain fragmented and reactive, with limited capacity for proactive, lifecycle-wide risk management. To address these issues, sustainability risks associated with AI are examined by analyzing key risk dimensions, evaluating current governance responses, and identifying directions for improvement. A structured review of literature and policy frameworks clarifies persistent limitations, including the absence of operational tools, inconsistent standards, and insufficient cross-sectoral coordination. To help close these gaps, the concept of an AI Green Index is proposed as an adaptable framework of traceable and comparable indicators, supporting the identification, assessment, and management of multidimensional AI sustainability risks for improved governance performance. By mapping these challenges, evaluating existing responses, and outlining concrete pathways forward, the work aims to inform policy design and promote the development of more effective, consistent, and inclusive strategies for sustainable AI governance.