Inventory of Reports on Sustainable AI
Explore a curated selection of reports on sustainable AI, published by our members. This page showcases key publications that shed light on the latest advances, challenges, and best practices in the design, deployment, and environmental impact of artificial intelligence.
The opinions expressed in these reports are those of the issuing organization and do not necessarily reflect the views of the Coalition’s members.
Key Reports & Publications on Green AI
Sustainable Procurement Guidelines for Data Centres and Servers
UNEP’s U4E developed these Guidelines to support government institutions and procurement authorities to establish clear, measurable criteria for improving the energy efficiency of both new and existing data infrastructure. The recommendations are based on key performance indicators such as power and water usage effectiveness, IT equipment efficiency, and cooling performance.
Frugal AI: Introduction, Concepts, Development and Open Questions
This report offers a comprehensive exploration of Frugal AI, highlighting its potential to drive cost-effective and sustainable innovation in resource-constrained settings. It examines the environmental footprint of AI technologies, underscores the need to optimize systems for efficiency and accessibility, and presents a range of strategies and open questions surrounding the development and deployment of Frugal AI.
ICT energy evolution: Telecom, data centers, and AI
This analysis examines electricity consumption within the ICT sector, revealing that although data traffic has surged exponentially, overall energy use has risen only modestly thanks to significant efficiency gains. The report also considers how future energy trends may be shaped by the rapid evolution and deployment of AI technologies.
Energy and AI
The rapid rise of artificial intelligence (AI) has raised critical questions about its impact on the energy sector. AI depends heavily on electricity especially for powering data centres yet it also holds potential to revolutionize how energy systems operate. Despite this dual role, policymakers and stakeholders have lacked the data needed to assess both its demands and benefits.
This International Energy Agency (IEA) report addresses that gap through new global and regional modelling, enriched by consultations with governments, regulators, tech companies, and energy experts. It projects AI’s electricity consumption over the next decade, identifies the energy sources likely to support it, and explores implications for energy security, emissions, innovation, and affordability.
Semiconductor Emission Explorer: Tracking Greenhouse Gas Emissions from Chip Production (2015-2023)
Analysis of greenhouse gas emissions from global chip manufacturers between 2015 and 2023, highlighting key trends and the need for improved reporting standards. This analysis examines greenhouse gas emissions from global chip manufacturers between 2015 and 2023, highlighting major trends and emphasizing the urgent need for more robust and standardized reporting practices.
Artificial intelligence, data, computation: what infrastructure for a low-carbon world?
This report investigates the growing environmental footprint of artificial intelligence and data infrastructure, with a particular focus on energy consumption and carbon emissions. As AI adoption accelerates, so too does the demand for electricity—especially in data centers that power large-scale model training and inference. The report highlights key trends, quantifies energy use across AI systems, and examines the implications for sustainability, emissions targets, and energy policy.
Power Hungry Processing: Watts Driving the Cost of AI Deployment?
In recent years, the rise of commercial AI products built on generative, multi-purpose systems has marked a shift toward a unified approach to integrating machine learning (ML) into technology. These models promise versatility across tasks, but this ambition of generality comes with significant environmental costs particularly in terms of energy consumption and carbon emissions.
Measuring the environmental impacts of artificial intelligence compute and applications
This report seeks to deepen understanding of AI’s environmental impacts and guide efforts to mitigate its negative effects while harnessing its potential to support planetary sustainability. It distinguishes between the direct impacts—such as energy use and emissions from developing, operating, and disposing of AI systems—and the indirect consequences of AI applications across sectors. Key recommendations include establishing standardized measurement frameworks, expanding data collection, identifying AI-specific environmental effects, considering impacts beyond operational energy use, and enhancing transparency and equity. These steps aim to equip policymakers to make AI a constructive force in addressing global sustainability challenges.
A Systemic Review of Green AI
As AI-based systems become increasingly widespread, their carbon footprint has grown too substantial to ignore. This has prompted calls for researchers and practitioners to take responsibility for the environmental impact of the models they develop and deploy. In response, a new field Green AI has emerged, dedicated to exploring the sustainability of artificial intelligence. Despite the surge in interest, a comprehensive synthesis of Green AI research remains lacking. To address this gap, this paper presents a systematic review of the existing literature on Green AI, aiming to map current efforts, identify key trends, and highlight opportunities for more sustainable AI development.
Artificial Intelligence and electricity: A system dynamics approach
This study offers an in-depth exploration of how artificial intelligence influences electricity consumption, examining a range of deployment scenarios from localized inference to large-scale model training and assessing the broader sustainability consequences. It considers both direct energy use in data centers and indirect effects across sectors, highlighting the trade-offs, opportunities, and policy considerations essential for aligning AI development with climate goals.
Navigating AI’s Thirst in a Water-Scarce World: A Governance Agenda for AI and the Environment
This report underscores the substantial water and energy requirements of artificial intelligence and data center infrastructure, drawing attention to the growing strain these technologies place on ecosystems and communities, especially amid escalating global water scarcity. It advocates for comprehensive environmental disclosure and robust governance frameworks to ensure transparency, accountability, and equitable resource management. By addressing both operational and lifecycle impacts, the report calls for urgent action to align AI development with planetary boundaries and sustainability goals.
Measuring and Standardizing AI’s Energy and Environmental Footprint to Accurately Assess Impacts
The rapid expansion of artificial intelligence (AI) is driving a surge in data center energy consumption, water use, carbon emissions, and electronic waste—yet these environmental impacts, and how they will evolve in the future, remain largely unclear. Without standardized metrics and reporting, policymakers and grid operators cannot accurately monitor or manage AI’s growing resource footprint.
This policy memo proposes a set of congressional and federal executive actions to establish comprehensive, standardized metrics for assessing AI’s energy and environmental impacts across model training, inference, and data center infrastructure.
Smarter, smaller, stronger: resource-efficient generative Al & the future of digital transformation
This report explores the dual role of artificial intelligence in environmental sustainability. While AI offers immense potential for reducing environmental impact through resource optimization and climate modeling, the rise of generative AI poses urgent challenges in terms of the consumption of energy, water and critical minerals. The report highlights the need to design AI systems “Clean by Design”, integrating energy and resource efficiency from the outset.
Measuring What Matters: How to Assess AI’s Environmental Impact
The report offers a comprehensive overview of current approaches to evaluating the environmental impacts of AI systems. The review focuses on identifying which components of AI’s environmental impacts are being measured, evaluating the transparency and methodology soundness of these measuring practices, and determining their relevance and actionability.
Artificial Intelligence:Generative AI’s Environmental and Human Effect
Generative AI could dramatically increase productivity and transform workloads in many industries. It can be used to respond to questions in customer service chats, create schedules, summarize information, produce Internet content—and more. But generative AI also poses potential risks to people and the environment. For example, the IT equipment that powers generative AI needs a lot of water and electricity to function efficiently and avoid overheating. Also, generative AI could replace workers or be used to create dangerous deepfakes. Our Technology Assessment discusses these and other challenges and offers options for policymakers to consider.
Key challenges for the environmental performance of AI
In the months leading up to the AI Action Summit, a wide range of stakeholders were consulted to elaborate collaboratively a position paper on the current key challenges to foster environmental performance and reduce the environmental impact of AI. The goal is to align internationally on the major challenges to achieve environmental performance of hardware and software and extend the lifespan of equipment and software used for AI.
Contribution to a global environmental standard for AI
Mistral AI has conducted a first-of-its-kind comprehensive study to quantify the environmental impacts of our LLMs. This report aims to provide a clear analysis of the environmental footprint of AI, contributing to set a new standard for our industry.
Generative AI Environmental Impact Study
This report allows to uncover AI’s carbon impact and strategies for sustainable innovation, by discovering best practices for balancing innovation and responsibility, understanding rising environmental demands and calculating the impact of two AI marketing use cases.
Small Language Models are the Future of Agentic AI
This research paper”Small Language Models are the Future of Agentic AI” argues that Small Language Models (SLMs) are better suited than Large Language Models (LLMs) for many agentic AI applications. It highlights that most agent tasks are repetitive and narrowly scoped, making SLMs more efficient, flexible, and cost-effective. The authors propose that SLMs can deliver sufficient reasoning and performance while reducing infrastructure and energy costs. They advocate for heterogeneous systems where SLMs handle routine tasks and LLMs are used selectively. A conversion algorithm from LLM-based agents to SLM-based ones is outlined to support adoption. The paper calls for broader discussion and contributions to promote sustainable and scalable AI agent design
Artificial Intelligence (AI) end-to-end: The Environmental Impact of the Full AI Lifecycle Needs to be Comprehensively Assessed – Issue Note
The United Nations Environment Programme (UNEP), as the leading global environmental authority, has been tasked by UN Member States to assess the environmental dimensions of digital technologies. This includes evaluating their potential to support sustainability and their environmental impacts. This note outlines key areas identified by UNEP regarding the environmental footprint of artificial intelligence (AI) across its lifecycle. It aims to inform Member States, civil society, the private sector, and the public, while encouraging the research community to develop scientific methods for objectively measuring AI’s environmental impact.
Measuring the environmental impact of delivering AI at Google Scale
This paper presents a comprehensive method for measuring the environmental impact of AI inference at Google scale. It shows that a single Gemini Apps text prompt uses just 0.24 Wh of energy far less than public estimates. Thanks to software optimizations and clean energy sourcing, Google reduced energy use by 33× and carbon footprint by 44× in one year. The authors advocate for standardized metrics to drive efficiency across the entire AI serving stack.
Key Reports & Publications on AI for Green
Artificial Intelligence for Climate Action:
Advancing Mitigation and Adaptation in Developing Countries
This technical paper offers a comprehensive overview for policymakers, practitioners, and researchers navigating opportunities, challenges, and risks of the use of AI for climate action in developing countries, with a focus on LDCs and SIDS as these countries face unique vulnerabilities to climate change. AI driven solutions can become potential enablers for adapting to climate impacts and reducing GHG emissions. However, risks and challenges also exist, which need to be addressed for the effective and sustainable use of AI in climate action.
Green and intelligent: the role of AI in the climate transition
Artificial Intelligence (AI) can play a powerful role in supporting climate action while boosting sustainable and inclusive economic growth. However, limited research exists on the potential influence of AI on the low-carbon transition. Here we identify five areas through which AI can help build an effective response to climate threats. We estimate the potential for greenhouse gas (GHG) emissions reductions through AI applications in three key sectors—power, food, and mobility—which collectively contribute nearly half of global emissions. This is compared with the increase in data centre-related emissions generated by all AI-related activities.
Leveraging artificial intelligence to enhance early action towards the Kunming-Montreal Global Biodiversity Framework
Developed on an on-demand basis through the Early Action Support (EAS) Project implemented by UNDP, NBSAP Target Similarity Assessments offer customized insights on synergies between global and national biodiversity targets and provide recommendations for enhanced alignment to bring about a transformation in our societies’ relationship with biodiversity by 2030. Developed with governments to address their unique needs, these assessments can foster dynamic, inclusive, and effective national stakeholder engagement to fill gaps, raise political will, and improve sectoral collaboration, resulting in accelerated progress towards global biodiversity commitments. This publication releases the methodology behind the National Biodiversity Strategies and Action Plans (NBSAP) Target Similarity Assessments and identifies key lessons learned and opportunities for future applications.
Tackling Climate Change with Machine Learning
Climate change is one of the greatest challenges facing humanity, and we, as machine learning (ML) experts, may wonder how we can help. Here we describe how ML can be a powerful tool in reducing greenhouse gas emissions and helping society adapt to a changing climate. From smart grids to disaster management, we identify high impact problems where existing gaps can be filled by ML, in collaboration with other fields. Our recommendations encompass exciting research questions as well as promising business opportunities. We call on the ML community to join the global effort against climate change.
AI for Ocean Conservation: The Role of Artificial Intelligence in Marine Sustainability
This paper explores the transformative role of AI in ocean conservation, highlighting its applications, benefits, and limitations. It emphasizes the necessity of integrating AI with existing marine conservation frameworks to enhance environmental sustainability, protect biodiversity, and mitigate the adverse effects of climate change on marine ecosystems..
Artificial Intelligence for Climate Action in Developing Countries: Opportunities,
Challenges and Risks
This information note has been prepared by the Technology Executive Committee (TEC) as part of the TEC rolling workplan 2023-2027. It provides an overview of the opportunities, risks and challenges of using artificial intelligence (AI) for climate action in developing countries, with a focus on least developed countries (LDCs) and small island developing States (SIDS).
Climate Change and AI: Recommendations for Government
The report highlights 48 specific recommendations for how governments can both support the application of AI to climate challenges and address the climate-related risks that AI poses. AI is already being used to support climate action in a wide range of use cases, several of which the report highlights. The authors also detail critical bottlenecks that are impeding faster adoption.
Aligning Artificial Intelligence with Climate Change Mitigation
Assessing and shaping the effects of artificial intelligence (AI) and machine learning (ML) on climate change mitigation demands a concerted effort across research, policy, and industry. However, there is great uncertainty regarding how ML may affect present and future greenhouse gas (GHG) emissions. This is owed in part to insufficient characterization of the different mechanisms through which such emissions impacts may occur, posing difficulties in measuring and forecasting them. We therefore introduce a systematic framework for describing ML’s effects on GHG emissions
Biodiversity and Artificial Intelligence: Opportunities & Recommendations for Action
Biodiversity loss is one of the most critical issues facing humanity, requiring urgent and coordinated action. Despite ongoing conservation efforts, biodiversity has declined dramatically in recent decades. Artificial intelligence (AI) is one tool that offers opportunities to accelerate action on biodiversity conservation. However, it must be deployed in a way that supports a paradigm shift to new, sustainable models of development, rather than entrenching business as usual. However AI is not asilver bullet and needs to be deployed as part of wider applications and efforts that support action.
UNDP Digital Guides : Nature
Signature Solution 4 – Environment
Digital technology plays a crucial role in achieving the UNDP Nature Pledge and the Kunming-Montreal Global Biodiversity Framework by enabling smart monitoring, data transparency, and traceability. However, it also risks widening inequalities, especially for vulnerable groups. UNDP promotes an inclusive digital transformation to harness technology for biodiversity while mitigating its negative impacts.
UNDP Digital Guides : Climate
Signature Solution 4 – Environment
Digital technologies offer powerful tools to mitigate and adapt to climate change, supporting better data, planning, and local engagement. UNDP helps over 120 countries integrate these tools into their climate strategies. However, the benefits are unevenly distributed, and risks like digital inequality, misinformation, and increased emissions must be addressed to ensure a just and inclusive transition.
The potential for AI to revolutionize conservation: a horizon scan
Artificial Intelligence (AI) is an emerging tool that could be leveraged to identify the effective conservation solutions demanded by the urgent biodiversity crisis. We present the results of our horizon scan of AI applications likely to significantly benefit biological conservation. An international panel of conservation scientists and AI experts identified 21 key ideas. These included species recognition to uncover dark diversity’, multimodal models to improve biodiversity loss predictions, moni toring wildlife trade, and addressing human–wildlife conflict.
AI for Climate and Nature: Landscape Assessment
Transformative system-oriented changes are urgently needed to confront the world’s most intractable challenges such as escalating biodiversity loss and climate change. Artificial intelligence (AI) provides a way to better understand these vital and complex problems and—if applied strategically—to dramatically accelerate the pace of solutions. This report examines how AI can help address these urgent challenges by enhancing our understanding and accelerating the implementation of impactful, scalable, and ethical solutions in climate and nature.
Pushing the frontiers in climate modelling and analysis with machine learning
Climate modelling and analysis are facing new demands to enhance projections and climate information. Here we argue that now is the time to push the frontiers of machine learning beyond state-of-the-art approaches, not only by developing machine-learning-based Earth system models with greater fidelity, but also by providing new capabilities through emulators for extreme event projections with large ensembles, enhanced detection and attribution methods for extreme events, and advanced climate model analysis and benchmarking. Utilizing this potential requires key machine learning challenges to be addressed, in particular generalization, uncertainty quantification, explainable artificial intelligence and causality.
Machine Learning for Sustainable Energy Systems
In recent years, machine learning has proven to be a powerful tool for deriving insights from data. In this review, we describe ways in which machine learning has been leveraged to facilitate the development and operation of sustainable energy systems. We first provide a taxonomy of machine learning paradigms and techniques, along with a discussion of their strengths and limitations. We then provide an overview of existing research using machine learning for sustainable energy production, delivery, and storage. Finally, we identify gaps in this literature, propose future research directions, and discuss important considerations for deployment.
ICEF Artificial Intelligence for Climate Change Mitigation Roadmap
The ICEF Artificial Intelligence for Climate Change Mitigation Roadmap (Second Edition) comprehensively updates all chapters in last year’s ICEF Roadmap (on topics including the power sector, food system and materials innovation) while adding six new chapters (on aviation, buildings, carbon capture, nuclear power, large language models and extreme weather response). Each chapter highlights 5-10 specific, actionable recommendations, providing the most detailed and comprehensive set of recommendations to date on how artificial intelligence (AI) can be used to respond to climate change.
Tackling Climate Change with Machine Learning
Climate change is one of the greatest challenges facing humanity, and we, as machine learning (ML) experts, may wonder how we can help. Here we describe how ML can be a powerful tool in reducing greenhouse gas emissions and helping society adapt to a changing climate. From smart grids to disaster management, we identify high impact problems where existing gaps can be filled by ML, in collaboration with other fields. Our recommendations encompass exciting research questions as well as promising business opportunities. We call on the ML community to join the global effort against climate change.
How AI Can Be a
Powerful Tool in the Fight
Against Climate Change
This report explores how advanced analytics and artificial intelligence (a pairing referred to the report simply as “AI”) are tools uniquely positioned to help manage the complex issues brought about by climate change. Three areas of application are particularly relevant: Mitigation, adaptation and resilience, and fundamentals.
List a new initiative
The initiatives identified as part of the coalition are brought forward by its members and supporters and listed in this “Hub”. They have to be open, collaborative and internationally-oriented. Members and supporters of the Coalition do not endorse the listed initiatives. However, they can participate or provide support to the listed initiatives at their own discretion. If you wish to list a new initiative, please use the contact form.
