Artificial Intelligence and Sustainability: A Conceptual Framework for System-Level Impact Assessment

Abstract

Artificial intelligence (AI) is rapidly emerging as a general-purpose technology with far-reaching implications for sustainable development. While AI applications are increasingly deployed across sectors such as healthcare, energy systems, urban management, and education, their overall sustainability impacts remain uncertain and often contradictory. Existing research typically examines isolated effects of AI within individual sustainability pillars, which limits the ability to understand systemic interactions, feedback loops, and long-term consequences. This study introduces a conceptual analytical framework designed to assess the sustainability impacts of artificial intelligence across environmental, economic, and social dimensions, extended by an additional individual-level pillar. The framework defines a set of AI Impact Groups (AIG) that translate technological capabilities into system-level functions, including perception, learning, strategic foresight, coordination, and risk detection. In addition, the model introduces key input parameters – AI intensity, adoption level, autonomy, quality of use, and system quality – that influence how AI capabilities translate into sustainability outcomes. By linking AI capabilities, system-level functions, and sustainability pillars, the proposed framework enables a more integrated assessment of both opportunities and risks associated with AI deployment. The model highlights how AI impacts propagate across domains and may generate both short-term benefits and long-term systemic risks, such as rebound effects, technological dependence, or skill erosion. The framework provides a foundation for future scenario analysis, sector-specific impact assessment, and interdisciplinary collaboration aimed at understanding and governing AI-driven sustainability transitions.