AI’s inextricable materiality is central to its environmental impact. From energy-intensive computation to the extractive demands of hardware and infrastructure, AI technologies are deeply embedded in global material and ecological systems. Examined through a socio-political lens, these environmental dimensions raise pressing justice concerns. This paper argues that discussions of AI’s environmental impacts must move beyond efficiency metrics to incorporate a global climate justice perspective within AI ethics, with particular attention to unequal global impacts. The paper unfolds in three sections. First, it demonstrates how AI’s environmental footprint across its full lifecycle, from development to deployment, constitutes a global climate justice concern. This includes analysis of energy use and water consumption in data centres, as well as the geopolitical implications of mineral extraction for hardware. Second, the paper examines what applying a climate justice perspective to AI ethics entails. Drawing on political philosophy, it focuses on two core dimensions of justice: distributional (who bears the costs and benefits of AI’s environmental impacts) and procedural (who is included in, or excluded from, decision-making about AI development and deployment). Together, these dimensions inform a set of conceptual and normative tools for assessing the AI-climate nexus, with particular relevance to Global South contexts. Finally, the paper grounds this discussion in a case study of India, illustrating how AI development intersects with climate vulnerability and global environmental inequality. The paper concludes by reflecting on how a climate justice approach can clarify the ethical stakes of AI’s environmental impacts and contribute to more inclusive and globally relevant AI governance.
Anand et al. (Sun,) studied this question.