carbon footprint estimation

Carbon emissions from generative AI inference have been systematically underestimated due to ambiguous system boundaries, lack of standardized metrics, and methodological biases in existing assessments. This study employs a scoping review to systematically analyze carbon footprint measurement tools, classification frameworks, and research trends across the full training–inference lifecycle. We propose (i) a standardized measurement protocol, (ii) a dynamic evaluation framework, and (iii) a multidimensional sustainability assessment system—novelly integrating user behavior and system boundary definitions into a unified analytical model. Our analysis uncovers nonlinear relationships between carbon emissions and model scale, prompt complexity, and deployment environment, identifying key leverage points for mitigation. The findings establish a reusable methodological benchmark and practical guidelines for AI environmental governance across technical, social, and operational dimensions.

This study addresses the fragmented understanding of environmental impacts across the full lifecycle of artificial intelligence systems, a gap marked by incomplete phase coverage, inconsistent metrics, and opaque methodologies in current “green AI” research. The authors propose a unified analytical framework encompassing eight stages—from hardware manufacturing and data processing to model training and deployment—and integrate life cycle assessment (LCA), systematic literature review, and multidimensional environmental indicators such as CO₂e emissions. Their analysis reveals that existing studies predominantly focus on training and inference while overlooking critical factors like water consumption, raw material extraction, and embodied carbon. To rectify these omissions, the paper advances a standardized, comparable, and policy-oriented pathway for assessing AI’s environmental footprint, thereby promoting more systematic and rigorous green AI research.

This study addresses the challenges of identifying software-related carbon emissions and clarifying responsibility boundaries in the ICT sector. We propose the first carbon accountability mapping framework spanning the entire software lifecycle. Leveraging systematic mapping, ICT energy consumption modeling, and multi-stakeholder governance analysis, we trace energy flows and emission sources across hardware-software interactions, and explicitly define climate accountability for developers, operators, and cloud service providers across design, deployment, and operation phases. Our key innovation lies in operationalizing abstract environmental responsibility into a traceable, allocatable role-based model, and developing a standardized Software Carbon Accountability Atlas tool. The framework supports compliance with EU regulatory frameworks—including the Corporate Sustainability Reporting Directive (CSRD) and Corporate Sustainability Due Diligence Directive (CSDD)—and significantly enhances transparency in ICT firms’ carbon disclosures and cross-organizational emission reduction coordination.

Scope 3 greenhouse gas emissions are challenging to analyze accurately due to sparse disclosure, heterogeneous formats, and a lack of traceable evidence. This work proposes the first dual-granularity framework for extracting Scope 3 emissions data at both organizational and building levels, featuring full traceability of supporting evidence. By integrating optical character recognition (OCR), large language models (LLMs), rule-based engines, and table reconstruction techniques, the framework enables end-to-end extraction of high-precision, interpretable Scope 1–3 emissions data from real-world ESG reports. Concurrently, the study introduces the first multimodal, evidence-annotated dataset designed to support reliable integration and transparent provenance tracking of heterogeneous emissions disclosures.

This work addresses the lack of a unified mechanism for tracking the environmental impact of open-source AI model derivatives—such as fine-tuned, quantized, or merged variants—rendering their energy use, water consumption, and carbon emissions largely invisible. To bridge this gap, the paper introduces the Data and Impact Accounting (DIA) framework, which extends environmental footprint tracking from base models to the entire lineage of derived models for the first time. DIA employs a lightweight, transparent layer to standardize carbon and water footprint metadata, integrates low-overhead measurement tools, and visualizes cumulative impacts through a public dashboard. By enabling comparability across model versions and lineages, DIA provides an unobtrusive, scalable infrastructure for sustainability accountability within the open-source AI ecosystem.

Addressing the challenge of imprecise software resource consumption measurement—which hinders green computing advancement—this paper proposes the Green Metrics Tool (GMT) framework. GMT establishes a controlled, reproducible, full-lifecycle measurement infrastructure within containerized environments, integrating multi-source resource monitoring and standardized metric collection. Methodologically, it innovatively combines Life Cycle Assessment (LCA) principles with large language models (LLMs) to enable visual energy consumption analysis, cross-project comparability, and intelligent optimization recommendations driven jointly by rule-based logic and semantic understanding. GMT significantly enhances automation and engineering practicality in green software evaluation, delivering actionable energy-efficiency diagnostics and optimization pathways for developers. Empirical validation demonstrates its effectiveness in identifying high-energy-consumption components and supporting sustainability-oriented improvements.

Assessing food carbon footprints remains challenging due to opaque global supply chains, fragmented data, and the methodological complexity of life cycle assessment (LCA). To address this, we propose a knowledge-enhanced AI framework that integrates LCA principles with retrieval-augmented generation (RAG) to build an interactive, explainable food carbon footprint analysis system. The system accepts arbitrary food items as input and supports multi-turn question-answering, dynamically translating technical emission data into intuitive analogies (e.g., “equivalent to driving X kilometers”). A publicly deployed prototype demonstrates feasibility in real-world environmental decision support. Empirical validation confirms improved interpretability and usability for non-expert stakeholders. However, limitations persist—including incomplete data coverage and ambiguities in system boundary definition. This work establishes a scalable technical pathway for promoting low-carbon food consumption, sustainable production practices, and evidence-informed climate policy.

Existing global greenhouse gas (GHG) emission datasets suffer from low accuracy, incomplete coverage, coarse spatiotemporal resolution, and significant reporting lags—impeding effective climate monitoring and evidence-based policymaking. To address these limitations, this study develops Climate TRACE, an open-access platform that delivers facility-level, monthly, near-real-time estimates of anthropogenic emissions across all major sectors—including power generation, industry, and transportation. Leveraging multi-source data—including satellite remote sensing, ground-based measurements, and legacy inventories—the platform employs sector-specific, AI-driven models to perform dynamic gap-filling and synergistic inverse modeling. Emission estimates are published monthly starting from January 2021, with a latency of only two months, and cover subnational administrative units worldwide. The platform is designed for accessibility by non-technical users. This work substantially enhances transparency, traceability, and responsiveness in emissions monitoring, establishing a new paradigm for data-driven climate governance.

This study addresses the widespread lack of rigor in model construction, calibration, and integration within current ICT life cycle assessment (LCA) practices, which undermines result credibility. To tackle this issue, the work proposes four foundational trustworthiness criteria for ICT LCA—model provenance, clearly defined scope, end-to-end traceability, and obsolescence management—and introduces an open computational LCA framework grounded in these principles. The framework integrates dependency graph modeling, a versioned model repository, automated enforcement of completeness constraints, and a structured model taxonomy. It effectively uncovers common misuses and structural flaws in existing LCA studies, thereby offering a viable pathway toward transparent, reproducible, and sustainably evolvable assessments of ICT environmental impacts.

This study addresses the limited scope of traditional high-performance computing (HPC) evaluations, which typically focus solely on performance and energy consumption while overlooking the comprehensive environmental costs of operational configurations. The authors propose the first job-level unified accounting framework that integrates both operational and full life-cycle (embodied) carbon and water footprints. Leveraging life-cycle assessment methodologies, real-time runtime monitoring, and hardware manufacturing emission data, the framework enables fine-grained quantification of environmental impacts. The analysis reveals that increasing thread count generally reduces total environmental footprints, albeit with diminishing marginal returns; while carbon footprints are predominantly driven by operational phases, water footprints are largely dominated by embodied impacts. By jointly incorporating both footprint types at the job granularity, this work establishes a novel paradigm for assessing HPC sustainability.

This work addresses the growing environmental impact of generative artificial intelligence (GenAI) in software development, where rising computational demands exacerbate energy consumption and carbon emissions, while existing governance mechanisms largely overlook sustainability concerns. To bridge this gap, the paper proposes a Carbon-Aware Governance Gateway (CAGG) architecture that uniquely integrates carbon budgeting and energy provenance tracking into the human-AI collaborative governance layer. The framework comprises three core components—an energy and carbon provenance ledger, a carbon budget manager, and a green validation orchestrator—enabling joint optimization of governance policies and sustainability objectives. CAGG facilitates low-carbon, transparent, and accountable GenAI development, significantly reducing the auxiliary energy use and carbon footprint of governance processes without compromising trustworthiness.

This study addresses the lack of standardized methodologies for Scope 3 Category 1 carbon accounting in enterprise AI inference services, where current practices often overestimate emissions by 10–40× due to reliance on coarse-grained industry-average emission factors. The authors propose the first four-tier hierarchical accounting framework aligned with CSRD compliance requirements, which degrades systematically—from a token-level physical energy consumption model to expenditure-driven environmentally extended input-output (EEIO) analysis—based on data availability. Integrating GPU energy benchmarks (ML.ENERGY v3), regional grid carbon intensities (EPA eGRID, Ember), and water efficiency metrics, the framework reveals a previously undocumented carbon-water trade-off: low-carbon regions may entail high water consumption, influencing data center siting decisions. An empirical assessment of a 200-person European firm shows annual AI inference emissions under 1 ton CO₂e, indicating that regulatory challenges stem from methodological gaps rather than emission magnitude, while also uncovering water-carbon synergies overlooked by mainstream ESG tools.