cs.SEMar 17, 2026

On the Carbon Footprint of Economic Research in the Age of Generative AI

Authors: Andres Alonso-Robisco, Carlos Esparcia, Francisco Jareño

Organizations: Banco de Espa˜na · Universidad de Castilla-La Mancha

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Generative artificial intelligence (AI) is increasingly used to write and refactor research code, expanding computational workflows. At the same time, Green AI research has largely measured the footprint of models rather than the downstream workflows in which GenAI is a tool. We shift the unit of analysis from models to workflows and treat prompts as decision policies that allocate discretion between researcher and system, governing what is executed and when iteration stops. We contribute in two ways. First, we map the recent Green AI literature into seven themes: training footprint is the largest cluster, while inference efficiency and system level optimisation are growing rapidly, alongside measurement protocols, green algorithms, governance, and security and efficiency trade-offs. Second, we benchmark a modern economic survey workflow, an LDA-based literature mapping implemented with GenAI assisted coding and executed in a fixed cloud notebook, measuring runtime and estimated CO2e with CodeCarbon. Injecting generic green language into prompts has no reliable effect, whereas operational constraints and decision rule prompts deliver large and stable footprint reductions while preserving decision equivalent topic outputs. The results identify human in the loop governance as a practical lever to align GenAI productivity with environmental efficiency.

Figures & tables

RQObjectiveMotivation
RQ1Quantify how alternative prompting strategies change the computational footprint of a decision-equivalent survey workflow, measured by runtime and CO 2 e.GenAI lowers the cost of producing and iterating code, which can increase throughput. The environmental cost of this productivity gain is not well measured at the workflow level, where repeated execution and redundant computation can arise.
RQ2Test whether generic green language is sufficient to reduce footprint, or whether operational constraints and explicit decision rules are required to deliver stable reductions in runtime and emissions.In practice, researchers often rely on high level instructions when interacting with LLMs. If such instructions do not reliably change what the system computes, then meaningful footprint reductions require governance through explicit constraints, not only normative appeals.
RQ3Assess practical output equivalence across prompting strategies by testing whether reduced footprint comes with any substantive change in topic model outputs.Footprint reductions are only meaningful if the analytical object is preserved. Because topic models are not identified across runs, we need a transparent equivalence check to separate genuine efficiency gains from changes in the underlying clustering of documents.
RQ4Conceptualise and test prompts as decision policies that allocate discretion between the researcher and the model, shaping the breadth of search and the amount of computation executed in AI-assisted research.The key margin is not whether an LLM is used, but how human machine interaction is structured. Prompt design can be treated as part of computational governance, since it determines stopping rules, search scope, and which outputs are computed and stored.
Table 1: Research questions, objectives, and motivations.
Figure 1: Identification and screening pipeline for the Green AI literature map. Counts reflect our arXiv retrieval and filtering protocol.
Figure 2: Model-selection diagnostics for the Green AI corpus: topic coherence ( cvc_{v} , higher is better) and held-out perplexity (lower is better) for LDA models estimated over K∈{5,…,15}K\in\{5,\ldots,15\} . Each point corresponds to one fitted model on the preprocessed abstract corpus; the main specification uses K=7K=7 based on the coherence plateau and interpretability checks.
TopicLabelShare of tokens (%)
1Carbon footprint and environmental impact of training28.1
2System-level energy optimisation (hardware and infrastructure)12.1
3Sustainability narratives, frameworks, and meta-research14.9
4Energy consumption and environmental methods8.8
5Inference and architecture optimisation for efficiency13.7
6Green algorithms and task specific learning11.2
Table 2: Green AI topic clusters from the K=7K=7 LDA model and their prevalence. “Share of tokens” reports the fraction of all preprocessed abstract tokens assigned to each topic under the fitted model (topics relabelled for presentation; see Appendix Table 8 ).
Figure 3: Topic co-occurrence in the Green AI corpus. Cells report the frequency with which two topics appear within the same document (based on document-level topic mixtures from the K=7K=7 LDA model), highlighting which themes tend to co-appear in abstracts.
Figure 4: Yearly evolution of Green AI themes. For each year, we count documents by their dominant LDA topic (highest topic share within the document) using the K=7K=7 model, showing how the composition of the retained arXiv corpus shifts over time.
TopicFocusRepresentative references
1Carbon footprint and environmental impact of training( Strubell et al. 2019 ; Henderson et al. 2020 ; Schwartz et al. 2019 )
2System-level energy optimisation (hardware and infrastructure)( Wang et al. 2025 ; Bian et al. 2023 ; Kadian and others 2025 )
3Sustainability narratives, frameworks, and meta-research( Schwartz et al. 2019 ; Verdecchia et al. 2023 ; Singh et al. 2024 )
4Energy consumption and environmental methods (measurement and reporting)( Lacoste et al. 2019 ; Courty et al. 2023 ; Henderson et al. 2020 )
5Inference and architecture optimisation for efficiency( Jegham et al. 2025 ; Li et al. 2024 ; Kadian and others 2025 )
6Green algorithms and task specific learning( Schwartz et al. 2019 ; Zhong et al. 2024 ; Ferdous et al. 2025 )
Table 3: Green AI topic clusters: focus and representative references (illustrative, not exhaustive).
Figure 5: Conceptual mechanism: prompt design allocates discretion between researcher and system, shaping search scope, stopping, and which outputs are computed, which determines executed compute and therefore runtime and CO 2 e. Output equivalence is required to interpret footprint reductions as efficiency gains.
Figure 6: Measurement boundary of the benchmark. We measure downstream workflow compute in Colab (runtime and CodeCarbon CO 2 e) and exclude upstream LLM inference for code generation/filtering as well as background machine activity unrelated to the instrumented workflow.
Figure 7: Schematic workflow of the LDA pipeline and measurement scope. Candidate topic counts are evaluated by fitting LDA models and computing coherence diagnostics; the stopping rule and the choice of K∗K^{\ast} depend on the prompting strategy (e.g. exhaustive grid, coarse search, or decision-driven early stopping). Measurement starts at preprocessing and covers model search/selection and post-estimation computation (runtime and CodeCarbon CO 2 e estimates).
Figure 8: Generic prompt used to generate a baseline LDA pipeline with AI-assisted coding.
Held constant in the experimentMay change in applied settingsExpected implication for footprint and interpretation
Task and workflow structure (LDA based survey pipeline)Other workflows (regressions, simulations, machine learning, data cleaning, writing with code execution, agentic pipelines)Absolute levels will differ, but the mechanism should strengthen when workflows involve wider search spaces and more optional outputs, since discretion has more scope to expand compute.
Execution environment (Google Colab free tier)Local machines, dedicated servers, cloud instances with different hardware and background loadBaseline runtime and emissions may shift, and noise may change, but decision rules should still reduce redundant execution when compute is unconstrained.
Model used for coding (ChatGPT 5.2)Other models and tools (different LLMs, code copilots, multi agent tools)More capable or more agentic systems can increase variance in compute when unconstrained. Governance via explicit rules should remain relevant as capability rises.
Measurement approach (runtime and CodeCarbon CO 2 e)Alternative carbon accounting assumptions, metered energy, different carbon intensity estimatesCO 2 e levels can vary with assumptions, which motivates reporting runtime alongside emissions. Relative differences across prompting strategies should be more stable than levels.
Table 4: Experimental controls and external validity considerations.
Figure 9: Stylised efficiency frontier for prompting strategies. The efficient frontier represents the maximum attainable breadth of exercise (practical scope) for a given computational cost C(p)C(p) (runtime or CO 2 e). Weak guidance (green soft) may leave outcomes below the frontier with little change relative to the naive baseline. Constraint-based prompting reduces cost but can also reduce breadth of exercise when the prompt narrows what is computed. Decision-driven prompting can further lower cost by restricting computation to decision-relevant outputs. Points are illustrative and not estimated.
Runtime
StrategyAbsolute savings (s)Relative (%)
Green soft1.030.05
Green computational constraints1011.0744.89
Decision-driven stewardship1251.3963.26
Carbon emissions
StrategyAbsolute savings (g CO 2 e)Relative (%)
Table 5: Average paired savings in runtime and estimated emissions relative to a within-block naive baseline ( n=5n=5 paired runs per strategy).
Figure 10: Absolute runtime (seconds) by paired repetition. Each panel shows a blocked comparison between a freshly re-run naive baseline (Naive) and the corresponding prompting strategy (Strategy) for n=5n=5 paired runs; lines connect within-pair outcomes.
Figure 11: Absolute emissions (g CO 2 e) by paired repetition. Each panel shows a blocked comparison between a freshly re-run naive baseline (Naive) and the corresponding prompting strategy (Strategy) for n=5n=5 paired runs; lines connect within-pair outcomes.
Comparison (vs. naive)Average matched Jaccard J‾\overline{J}
Green soft1.000
Green computational constraints1.000
Decision-driven (chosen K∗=7K^{\ast}=7 )1.000
Table 6: Sanity check of topic-output consistency relative to the naive baseline. Topics are compared at K=7K=7 using the top 10 words per topic. Because topic labels are not identified across runs, topics are aligned using the Hungarian algorithm to maximise total similarity across topics. Similarity is measured by the Jaccard index over sets of top words. Under fixed seeds and identical specifications, J‾=1\overline{J}=1 is expected and is interpreted as a verification that the analytical pipeline is unchanged across prompting strategies.
 Productivity (research output)Efficiency (computational footprint)
Primary marginLower time and cognitive costs for producing, refactoring, and debugging code; faster iterationRuntime, energy use, and CO 2 e of executed workflows, including optional search and post-estimation outputs
What GenAI changesReduces the marginal cost of iteration and enables more rapid prototyping of survey pipelinesCan widen the realised search scope and expand computed artefacts when discretion is not constrained
Main opportunityFaster construction and updating of computational surveys under scaleScope to reduce avoidable computation through explicit decision rules and stopping criteria
Main risk (if unmanaged)“More iterations” may shift effort toward exploration and complexity without proportional analytical gains“More automation” may translate into redundant execution, inflated runtime, and higher emissions
What our benchmark suggestsPrompting strategies preserve the ability to generate end-to-end pipelinesGeneric green language has limited and unstable effects; operational constraints and decision-driven prompts can materially reduce runtime and estimated CO 2 e in this workflow
Role of human judgementShapes research direction, interpretation, and what is read in depthEncodes search scope, stopping behaviour, and which outputs are decision-relevant (computational governance)
Table 7: GenAI-assisted survey workflows: productivity gains and computational-efficiency risks.
Appendix figures & tables9 assets

Supplementary material from the paper’s appendix.

Appendix

Figure 12: Yearly evolution of the retained Green AI corpus. Bars report the number of arXiv records classified as Green AI (titles and abstracts, after filtering) by publication year, illustrating the recent acceleration in relevant preprints.
Internal IDPresentation TopicMatching cues (high-probability tokens)
01model, training, carbon, footprint, performance
12energy, system, network, efficiency, consumption
23framework, research, paper, sustainable, ai
34energy, consumption, environmental, method, efficiency
45inference, deep, accuracy, computational, task
56task, learning, efficiency, computational, performance
Appendix
Table 8: Mapping between internal LDA topic IDs (0–6) and presentation topics (1–7).
TopicTop tokens with probabilities
10.024 model , 0.016 ai , 0.010 training , 0.010 carbon , 0.009 learning , 0.007 system , 0.007 performance , 0.007 footprint , 0.006 data , 0.006 computational
20.015 energy , 0.011 system , 0.010 ai , 0.008 model , 0.007 performance , 0.007 network , 0.006 efficiency , 0.006 consumption , 0.006 data , 0.005 learning
30.017 ai , 0.009 data , 0.008 energy , 0.006 system , 0.006 network , 0.006 framework , 0.005 research , 0.005 paper , 0.005 approach , 0.005 sustainable
40.019 ai , 0.017 model , 0.013 energy , 0.009 data , 0.006 efficiency , 0.006 consumption , 0.005 learning , 0.005 environmental , 0.005 approach , 0.005 method
50.018 model , 0.009 network , 0.009 inference , 0.008 deep , 0.007 learning , 0.007 accuracy , 0.007 computational , 0.007 task , 0.007 energy , 0.006 performance
60.014 energy , 0.012 model , 0.007 efficiency , 0.007 learning , 0.007 task , 0.005 network , 0.005 computational , 0.005 performance , 0.005 consumption , 0.004 accuracy
Appendix
Table 9: Top tokens per topic and estimated probabilities p(w∣k)p(w\mid k) in the final K=7K=7 model. Topics are ordered according to the presentation mapping in Table 8 .
Figure 13: Topic co-occurrence network for the K=7K=7 Green AI map. Nodes are topics and edge weights reflect document-level co-occurrence frequency (topics that jointly receive non-trivial weight within the same abstract), offering a graph view of the heatmap in Figure 3 .
Figure 14: Word clouds for the K=7K=7 Green AI topics. Word size is proportional to the estimated topic–word probability p(w∣k)p(w\mid k) from the LDA model fitted on preprocessed abstracts (lowercased, stopwords removed, lemmatised). Topic numbering follows the presentation mapping in Table 8 .
Strategy vs. naiveΔ‾T\overline{\Delta}_{T}sΔTs_{\Delta_{T}}ttpp
Green soft1.0328.830.080.940
Green computational constraints1011.078.11278.85<0.001<0.001
Decision-driven stewardship1251.3950.0055.97<0.001<0.001
Appendix
Table 10: Paired tt -tests relative to the naive baseline ( n=5n=5 , df=4df=4 ): runtime differences (seconds).
Strategy vs. naiveΔ‾E\overline{\Delta}_{E}sΔEs_{\Delta_{E}}ttpp
Green soft0.0100.0630.360.734
Green computational constraints5.1490.041278.81<0.001<0.001
Decision-driven stewardship2.5320.10156.04<0.001<0.001
Appendix
Table 11: Paired tt -tests relative to the naive baseline ( n=5n=5 , df=4df=4 ): emissions differences (grams CO 2 e).
StrategyMeanMedianBoot CI95 (mean)Boot CI95 (median)Pos.pp
Soft1.032.98[-23.83, 20.40][-44.99, 31.33]3/51.0000
Comp. constraints1011.071011.16[1004.62, 1017.11][999.72, 1020.54]5/50.0625
Decision-driven1251.391244.57[1214.45, 1292.22][1197.23, 1323.99]5/50.0625
Appendix
Table 12: Robustness checks for paired runtime savings relative to the within-block naive baseline ( n=5n=5 paired repetitions). We report mean and median paired savings, non-parametric bootstrap 95% confidence intervals (percentile, B=20,000B=20{,}000 resamples), and an exact two-sided sign test based on the number of positive paired savings.
StrategyMeanMedianBoot CI95 (mean)Boot CI95 (median)Pos.pp
Soft0.0100.007[-0.040, 0.055][-0.086, 0.071]3/51.0000
Comp. constraints5.1495.150[5.116, 5.180][5.091, 5.197]5/50.0625
Decision-driven2.5322.518[2.457, 2.615][2.423, 2.679]5/50.0625
Appendix
Table 13: Robustness checks for paired emissions savings relative to the within-block naive baseline ( n=5n=5 paired repetitions). We report mean and median paired savings, non-parametric bootstrap 95% confidence intervals (percentile, B=20,000B=20{,}000 resamples), and an exact two-sided sign test based on the number of positive paired savings. Emissions are expressed in grams of CO 2 e.
More in this topic

AI-Assisted Scientific Research

AI-assisted scientific research uses machine learning and AI agents to support discovery, analysis, experimentation, and scientific communication.

See all 251 papers

Explore similar work

CardsList
Apr 5, 2026

Toward a Sustainable Software Architecture Community: Evaluating ICSA's Environmental Impact

Generative AI (GenAI) tools are increasingly integrated into software architecture research, yet the environmental impact of their computational usage remains largely undocumented. This study presents the first systematic audit of the carbon footprint of both the digital footprint from GenAI usage in research papers, and the traditional footprint from conference activities within the context of the IEEE International Conference on Software Architecture (ICSA). We report two separate carbon inventories relevant to the software architecture research community: i) an exploratory estimate of the footprint of GenAI inference usage associated with accepted papers within a research-artifact boundary, and ii) the conference attendance and operations footprint of ICSA 2025 (travel, accommodation, catering, venue energy, and materials) within the conference time boundary. These two inventories, with different system boundaries and completeness, support transparency and community reflection. We discuss implications for sustainable software architecture, including recommendations for transparency, greener conference planning, and improved energy efficiency in GenAI operations. Our work supports a more climate-conscious research culture within the ICSA community and beyond

Mahyar T. Moghaddam, Mina Alipour, Torben Worm +1

SDU Software Engineering Odense, Denmark

Responsible AISoftware Engineering
AskPDF
Sep 26, 2026

Environmental Impact of Generative and Agentic AI: An in-Depth Analysis and Green Solutions

The proliferation of generative and agentic artificial intelligence (AI) systems has introduced computational demands whose environmental consequences are substantial yet underexamined. This paper examines the environmental footprint of modern AI systems across energy consumption, carbon emissions, water usage, and electronic waste over the full lifecycle of large language models, multimodal foundation models, and agentic workflows, from hardware fabrication and training through fine-tuning and inference to end-of-life disposal. This work provides a conceptual analysis, utilizing order-of-magnitude estimations based on published data, without conducting original physical measurements. We contribute a lifecycle taxonomy that crosses lifecycle phases with five impact dimensions and emission scopes; a Sustainability Assessment Framework for AI Systems (SAFIA) comprising nine indicators; a comparative analysis of traditional, generative, and agentic AI; seven open challenges; policy recommendations for regulators, cloud providers, hardware manufacturers, and AI developers; and a research roadmap to 2035. The analysis indicates that inference can rival or exceed training energy over a deployment lifetime, that agentic workflows can multiply the energy of equivalent single-pass inference by one to several orders of magnitude depending on the number of model and tool calls, that indirect water use and embodied carbon are systematically underreported, and that measurement tools, disclosure practices, and regulation have not kept pace with agentic deployment. These findings call for agent-aware energy measurement, lifecycle-based carbon and water accounting, and mandatory disclosure for large-scale AI training and deployment.

Abderaouf Bahi, Amel Ourici, Ibtissem Gasmi

Computer Science and Applied Mathematics Laboratory (LIMA), Faculty of Science and Technology, Chadli Bendjedid University, P.O. Box 73, El Tarf 36000, Algeria · Mathematical Modeling and Numerical Simulation Laboratory (LAM2SIN), Faculty of Technology, Badji Mokhtar University, P.O. Box 12, Annaba 23000, Algeria

Life Cycle AssessmentResponsible AI
AskPDF
Apr 13, 2026

Environmental Footprint of GenAI Research: Insights from the Moshi Foundation Model

New multi-modal large language models (MLLMs) are continuously being trained and deployed, following rapid development cycles. This generative AI frenzy is driving steady increases in energy consumption, greenhouse gas emissions, and a plethora of other environmental impacts linked to datacenter construction and hardware manufacturing. Mitigating the environmental consequences of GenAI remains challenging due to an overall lack of transparency by the main actors in the field. Even when the environmental impacts of specific models are mentioned, they are typically restricted to the carbon footprint of the final training run, omitting the research and development stages. In this work, we explore the impact of GenAI research through a fine-grained analysis of the compute spent to create Moshi, a 7B-parameter speech-text foundation model for real-time dialogue developed by Kyutai, a leading privately funded open science AI lab. For the first time, our study dives into the anatomy of compute-intensive MLLM research, quantifying the GPU-time invested in specific model components and training phases, as well as early experimental stages, failed training runs, debugging, and ablation studies. Additionally, we assess the environmental impacts of creating Moshi from beginning to end using a life cycle assessment methodology: we quantify energy and water consumption, greenhouse gas emissions, and mineral resource depletion associated with the production and use of datacenter hardware. Our detailed analysis allows us to provide actionable guidelines to reduce compute usage and environmental impacts of MLLM research, paving the way for more sustainable AI research.

Marta López-Rauhut, Loic Landrieu, Mathieu Aubry +1

LIGM, CNRS, Univ Gustave Eiffel, ENPC, Institut Polytechnique de Paris, France · Université Paris-Saclay, CNRS, LISN, Orsay, France

Life Cycle AssessmentResponsible AIEnergy-Efficient ML
AskPDF