Artificial intelligence (AI) is becoming an increasingly important factor in the climate transition. AI can improve energy efficiency, optimise resource use, accelerate climate innovation and strengthen climate forecasting, adaptation and risk management. At the same time, the development and deployment of AI require substantial computational resources, increasing electricity demand and, depending on the carbon intensity and timing of electricity use, associated emissions. We examine AI’s direct environmental footprint, its contribution to climate mitigation and adaptation, and rebound effects arising from lower costs, higher productivity and wider adoption. The overall impact of AI on climate change also depends on how AI capabilities evolve. We consider two stylised scenarios. Under an AI copilot scenario, benefits and environmental costs are likely to build up more gradually, but the net climate effect remains uncertain, depending on whether efficiency gains outpace growth in AI use, on the carbon intensity of additional electricity needed to power AI, and on how productivity gains translate into economic activity and resource use. Under a more transformative scenario involving artificial general intelligence (AGI), with cognitive capabilities comparable to those of humans across a broad range of tasks, both the potential benefits and environmental risks could become larger and more uncertain. The paper also reviews emerging policy responses and highlights the relevance of the AI–climate nexus for central banks. By affecting productivity, energy systems and climate-related risks, the AI–climate nexus can influence potential output and inflation dynamics, while the scale and financing of investment in AI and energy infrastructure may also have implications for financial stability.
