As machine learning (ML) continues its rapid expansion, the environmental cost of model training and inference has become a critical societal concern. Existing benchmarks overwhelmingly focus on standard performance metrics such as accuracy, BLEU, or mAP, while largely ignoring energy consumption and carbon emissions. This single-objective evaluation paradigm is increasingly misaligned with the practical requirements of large-scale deployment, particularly in energy-constrained environments such as mobile devices, developing regions, and climate-aware enterprises. In this paper, we propose AI-CARE, an evaluation tool for reporting energy consumption, and carbon emissions of ML models. In addition, we introduce the carbon-performance tradeoff curve, an interpretable tool that visualizes the Pareto frontier between performance and carbon cost. We demonstrate, through theoretical analysis and empirical validation on representative ML workloads, that carbon-aware benchmarking changes the relative ranking of models and encourages architectures that are simultaneously accurate and environmentally responsible. Our proposal aims to shift the research community toward transparent, multi-objective evaluation and align ML progress with global sustainability goals. The tool and documentation are available at https://github.com/USD-AI-ResearchLab/ai-care.
Figures & tables
Fig. 1 : Sequence-style control-flow diagram of the proposed reporting framework. The main process spawns external monitoring and carbon-emission services and executes models on the evaluation dataset. During execution, energy consumption E(m) is measured periodically, while carbon-emission values are retrieved asynchronously. After execution completes, performance, energy, and carbon metrics are normalized and aggregated to generate final reports. The framework operates purely as a reporting layer and does not influence model execution, training, or optimization.
Fig. 2 : Global carbon–performance tradeoff across all evaluated model–dataset pairs. Each subplot reports task performance (Accuracy, Precision, Recall, and F1-Score) versus total carbon emissions (training + inference) shown on a logarithmic scale. Colors indicate datasets, while marker shapes denote model families.
Fig. 3 : Metric-wise scalar carbon-aware scores across datasets and model families. Rows correspond to evaluation metrics (accuracy, precision, recall, and F1-score), while columns correspond to datasets (MNIST, Fashion-MNIST, CIFAR-10, CIFAR-100, and ImageNet-100). Each score integrates normalized task performance with total carbon emissions (training + inference). The grouped layout enables direct cross-metric and cross-dataset comparison of carbon-aware model rankings.
As the capabilities and ubiquity of Large Language Models (LLMs) grow, so does their environmental footprint. Despite calls for responsible AI, the machine learning community lacks standardised practices for carbon accounting. Our automated literature review of the 5,285 papers accepted to NeurIPS 2025 reveals that reporting of environmental impact is nearly non-existent. To catalyse a shift toward sustainable AI, we define standardised sustainability metrics for evaluating model training efficiency, accompanied by simple heuristics to estimate the carbon cost of LLM inference. We implement these metrics in carbonbenchmark, a drop-in software solution for tracking and reporting emissions. Finally, to combat the pursuit of marginal accuracy gains at disproportionate environmental costs, we formalise the Smallest Model that Achieves the Job' (SMAJ), a framework which challenges the field to prioritise computational efficiency and environmental accountability alongside traditional State-of-the-Art' (SotA) accuracy.
Lachlan McGinness, Dan Pagendam, Robert Offner
School of Computing, Australian National University · Commonwealth Scientific and Industrial Research Organisation
As machine learning and artificial intelligence find their way into nearly every aspect of climate, weather, and Earth system modeling, it is worth pausing to consider what our design decisions imply for the science and for the computational resources we consume. A growing body of literature addresses the ethical and sustainable development of ML/AI, yet translating these principles into day-to-day research practice remains a challenge as most of best practices are dispersed across multiple studies and commentaries. Here, we distill these discussions into a practical checklist that ML/AI and Earth system science practitioners can use to assess and reduce the environmental footprint of their own applications, organised around the successive stages of the model development pipeline. We complement the checklist with a selection of metrics drawn from the literature for estimating the energy consumption and carbon footprint of a project. For each question, we point to concrete examples and actionable suggestions from recent literature, aiming to bridge the gap between aspirational principles and the decisions researchers face at every stage of the development cycle.
The carbon footprint of any deployed Large Language Model (LLM) accumulates during inference, where repeated use of the model substantially exceeds the one-time cost of fine-tuning. Yet most efficiency interventions target either pre-training scale or post-hoc compression. We ask whether folding a calibrated, differentiable energy surrogate into the fine-tuning objective can produce inference behavior that gains task accuracy at zero or near-zero carbon cost, a break-even configuration. We propose a joint loss mechanism with a per-model carbon-emission parameter, a linear surrogate over parameter norm, FLOP proxy, and a memory proxy, fit from on-hardware energy profiling. We fine-tune three architecturally distinct families: Gemma-2 2B, Llama-3.1 8B, and Qwen-2.5 14B, and evaluate inference F1 and CO2 emissions on three MMLU subjects: abstract algebra, philosophy, and formal logic. We discover from several outcomes that the carbon term behaves as either harmful interference or beneficial regularization depending on the task structure. We position calibrated carbon-aware fine-tuning as a lightweight, drop-in regularizer with a non-empty but model and task-dependent break-even region. This is an ongoing work, and we will release our codebase soon.