← Back to all news
arXiv AI

Beyond State-of-the-Art: Standardising Environmental Impact Metrics for AI Research

The paper highlights that as Large Language Models grow in capability and prevalence, their environmental footprint is increasing, yet the machine learning community lacks standardized carbon accounting practices. An automated review of 5,285 NeurIPS 2025 papers shows almost no reporting of environmental impact. To address this, the authors propose standardized sustainability metrics for training efficiency, heuristics for estimating inference carbon costs, a software tool called carbonbenchmark for tracking emissions, and the SMAJ framework to encourage prioritizing computational efficiency and environmental accountability over marginal accuracy gains.

Related stories

arXiv Machine Learning
Sep 18

The Environmental Impacts of Language Model Training Keep Rising Now is the Time to Catch Impacts on the Rebound

The paper analyzes the environmental footprint of machine learning model training, focusing on large language models and their hardware. It finds that energy use and environmental impacts have risen exponentially over the past decade, even when employing carbon‑efficient electricity and more efficient hardware. The study argues that optimization strategies alone cannot curb these impacts due to a rebound effect, and stresses the need to evaluate hardware life‑cycle impacts and integrate environmental metrics into NLP research practices.

By Cl\'ement Morand (STL), Anne-Laure Ligozat (ENSIIE, LISN, STL), Aur\'elie N\'ev\'eol (STL, LISN)
llmsnlpbenchmarks
More like this →
arXiv AI
Jun 11

The Environmental Cost of LLMs in AIED: Reporting and Practices

arXiv:2606. 11215v1 Announce Type: cross Abstract: Large Language Model (LLM) usage in recent years has become increasingly widespread in the Artificial Intelligence in Education (AIED) community.

By Sabrina C. Eimler, Lukas Erle, Daniel Flood, Aditi Haiman, Luca H\"ackert, Andr\'e Helgert, Lachlan McGinness, B\"usra Yapici
llms
More like this →
arXiv AI
Jun 10

Accounting for AI Inference in Corporate GHG Inventories: A Four-Tier Methodology for Scope 3 Category 1 Reporting

arXiv:2606. 10660v1 Announce Type: cross Abstract: AI inference services -- API subscriptions, enterprise chat tools, and SaaS products with embedded AI features -- fall unambiguously within Scope 3 Category 1 under the Corporate Sustainability Reporting Directive (CSRD), which requires disclosure for fiscal years starting January 2024.

By Guillermo Llopis (SOMA AI, Barcelona)
benchmarks
More like this →
arXiv Machine Learning
Sep 3

Toward Explainable and Policy-Aware AI for Carbon Credit Price Prediction: A Research Framework for Emerging Carbon Markets

The paper proposes EPA-CarbonNet, a six‑layer model that fuses carbon market price series with policy text via cross‑attention and calibrated intervals, aiming to provide explainable, policy‑aware predictions for carbon credit prices. It evaluates the approach on eleven years of daily S&P carbon index data, finding that a simple random walk outperforms the model on five‑day RMSE, while the model achieves the best directional accuracy at 58.6%. The study identifies ten recurring gaps in current research and releases all code, data, and results publicly.

By Summaiya Unnisa Begum, Mohammed Nadeem Ullah, Mohammed Abdul Ghani Khan
More like this →
arXiv Machine Learning
Aug 19

Wasted large language models: A life cycle thinking approach

The paper argues that large language models (LLMs) have a growing carbon footprint and that current efficiency gains are offset by rebound effects such as Jevons Paradox. It proposes applying the EU waste hierarchy—prevention, reuse, recycling, recovery, and disposal—to LLMs, suggesting that preventing waste and unnecessary use can significantly reduce environmental impact. The authors emphasize that treating LLMs as products that can become waste offers new ways to motivate more sustainable practices.

By Erik Johannes Husom, Maria Emine Nylund, Ophelia Prillard
llms
More like this →
arXiv Machine Learning
Aug 14

Performance-Carbon Trade-Offs across Architectural Biases in Shear Flow Forecasting

arXiv:2509. 24517v3 Announce Type: replace Abstract: Development of modern deep learning methods has been driven primarily by the push for improving model efficacy (accuracy metrics), leading to large-scale models that require massive computational resources and result in considerable carbon footprint across the model lifecycle.

By Sophia N. Wilson, Jens Hesselbjerg Christensen, Raghavendra Selvan
More like this →
arXiv Machine Learning
Jul 13

A Survey on the Green Development of Large Models: From Resource-Efficient Architectures to Hardware-Software Co-Design

arXiv:2607. 09084v1 Announce Type: new Abstract: The rapid expansion of large-scale AI models has led to significant performance breakthroughs across diverse domains, yet it has also raised critical concerns regarding computational costs, energy consumption, and environmental sustainability.

By Linhui Xiao, Guiping Cao, Mingyue Guo, Xianchao Guan, Fan Yang, Ming Tao, Xin Li, Yuxin Peng, Yaowei Wang
fine-tuning
More like this →