The field of artificial intelligence and computing is moving towards a more sustainable and energy-efficient direction. Researchers are exploring ways to reduce the environmental impact of AI training and deployment, such as optimizing hardware and software co-design, improving data-loading latency and energy consumption, and developing carbon-aware execution methods. These innovative approaches aim to minimize the carbon footprint of AI systems without compromising performance. Notable papers in this area include: EMLIO, which introduces an efficient machine learning I/O service that minimizes end-to-end data-loading latency and I/O energy consumption. Sustainable AI Training via Hardware-Software Co-Design, which explores environmentally driven performance optimization methods for advanced GPU architectures. Multi-Metric Algorithmic Complexity, which proposes a weighted-operation complexity model that assigns realistic cost values to different instruction types across multiple dimensions. A Systematic Evaluation of the Potential of Carbon-Aware Execution for Scientific Workflows, which shows the potential for carbon-aware workflow execution and estimates the carbon footprint of real-world workflows. Measuring the environmental impact of delivering AI at Google Scale, which proposes a comprehensive methodology for measuring the energy usage, carbon emissions, and water consumption of AI inference workloads in a large-scale production environment.