{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,14]],"date-time":"2026-10-14T07:44:00Z","timestamp":1791963840590,"version":"4.3.4"},"reference-count":16,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["JMIR AI"],"abstract":"Abstract<\/jats:title>\n \n Background<\/jats:title>\n AI research increasingly depends on energy-intensive computation, yet energy use, greenhouse gas emissions, hardware life cycle burdens, and water consumption are rarely reported in a standardized way. This lack of reproducible environmental accounting limits comparisons across studies and obscures the trade-offs among model performance, infrastructure choices, carbon intensity, and cooling water demand.<\/jats:p>\n <\/jats:sec>\n \n Objective<\/jats:title>\n This study aimed to develop and describe an open-access, life cycle\u2013informed AI Environmental Footprint Calculator and propose a minimum reporting dataset for transparent environmental disclosure in AI research.<\/jats:p>\n <\/jats:sec>\n \n Methods<\/jats:title>\n We developed a browser-based calculator that combines operational energy, regional grid carbon intensity, power usage effectiveness (PUE), water usage effectiveness, hardware embodied emissions, and workload-specific metrics for training and inference. The framework was evaluated in 3 representative scenarios: a single\u2013graphics processing unit laboratory fine-tuning task, a midsized academic cluster workload, and a large-scale industrial training cycle. Outputs were compared with those of established tools to identify how boundary choices and parameter assumptions affect emission estimates.<\/jats:p>\n <\/jats:sec>\n \n Results<\/jats:title>\n \n Across the scenarios, the inclusion of PUE and hardware life cycle allocation increased reported emissions compared with operational-only estimates. In the small laboratory scenario, optimization reduced total emissions from approximately 0.07 kg carbon dioxide equivalents (CO\n 2<\/jats:sub>\n e) to 0.05 kg CO\n 2<\/jats:sub>\n e and improved the proposed label from C to B. In the midsized cluster scenario, carbon-aware scheduling reduced emissions from approximately 665 kg CO\n 2<\/jats:sub>\n e\/month to 450 kg CO\n 2<\/jats:sub>\n e\/month\u2014a 32% reduction. In the large-scale scenario, shifting to renewable-backed, lower-PUE infrastructure reduced operational emissions by approximately 79% while increasing the importance of water-carbon trade-off reporting.\n <\/jats:p>\n <\/jats:sec>\n \n Conclusions<\/jats:title>\n The calculator provides a practical and transparent method for reporting AI environmental footprints using auditable parameters and publication-ready outputs. Routine disclosure of energy use, emissions, water use, hardware assumptions, and regional context can improve reproducibility and support more equitable and sustainable AI research evaluation.<\/jats:p>\n <\/jats:sec>","DOI":"10.2196\/90770","type":"journal-article","created":{"date-parts":[[2026,5,18]],"date-time":"2026-05-18T14:50:06Z","timestamp":1779115806000},"page":"e90770-e90770","source":"Crossref","is-referenced-by-count":2,"title":["Quantifying and Disclosing the Environmental Footprint of AI in Research: Life Cycle\u2013Informed Framework and Open-Access Calculator Development Study"],"prefix":"10.2196","volume":"5","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-8913-9049","authenticated-orcid":false,"given":"Kaveh","family":"Mozafari","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4048-1705","authenticated-orcid":false,"given":"Yuanchao","family":"Ma","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-4589-6510","authenticated-orcid":false,"given":"Mohsen","family":"Amoei","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1273-9393","authenticated-orcid":false,"given":"Bertrand","family":"Lebouche","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5590-8866","authenticated-orcid":false,"given":"Esli","family":"Osmanlliu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6267-6140","authenticated-orcid":false,"given":"Dan","family":"Poenaru","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","published-online":{"date-parts":[[2026,8,18]]},"reference":[{"key":"R1","unstructured":"Schneider I Xu H Benecke S et al. Life-cycle emissions of AI hardware: a cradle-to-grave approach and generational trends. arXiv. Preprint posted online on Feb 1, 2025. [doi: 10.48550\/arXiv.2502.01671]"},{"key":"R2","unstructured":"Morrison J Na C Fernandez J Dettmers T Strubell E Dodge J . Holistically evaluating the environmental impact of creating language models. Presented at: International Conference on Learning Representations 2025 (ICLR 2025); Apr 24-28, 2025; Singapore. URL: https:\/\/openreview.net\/pdf?id=04qx93Viwj [Accessed 30-06-2026]"},{"key":"R3","unstructured":"Jegham N Abdelatti M Koh CY Elmoubarki L Hendawi A . How hungry is AI? benchmarking energy, water, and carbon footprint of LLM inference. arXiv. Preprint posted online on May 14, 2025. [doi: 10.48550\/arXiv.2505.09598]"},{"issue":"8062","key":"R4","doi-asserted-by":"publisher","first-page":"331","DOI":"10.1038\/s41586-025-08832-3","article-title":"Using life cycle assessment to drive innovation for sustainable cool clouds","volume":"641","author":"Alissa","journal-title":"Nature"},{"key":"R5","unstructured":"Desroches C Chauvin M Ladan L Vateau C Gosset S Cordier P . Exploring the sustainable scaling of AI dilemma: a projective study of corporations\u2019 AI environmental impacts. arXiv. Preprint posted online on Jan 24, 2025. [doi: 10.48550\/arXiv.2501.14334]"},{"key":"R6","doi-asserted-by":"publisher","DOI":"10.1016\/j.eiar.2026.108525","article-title":"More than carbon: cradle-to-grave environmental impacts of GenAI training on the Nvidia A100 GPU","volume":"121","author":"Falk","journal-title":"Environ Impact Assess Rev"},{"key":"R7","unstructured":"Energy and AI. International Energy Agency (IEA); 2025. URL: https:\/\/iea.blob.core.windows.net\/assets\/de9dea13-b07d-42c5-a398-d1b3ae17d866\/EnergyandAI.pdf [Accessed 30-06-2026]"},{"key":"R8","unstructured":"Jhaveri M Palat V . Measuring and standardizing AI\u2019s energy and environmental footprint to accurately access impacts. Federation of American Scientists. Jun 27, 2025. URL: https:\/\/fas.org\/publication\/measuring-and-standardizing-ais-energy-footprint\/ [Accessed 30-07-2026]"},{"key":"R9","doi-asserted-by":"publisher","DOI":"10.1016\/j.procir.2025.01.079","article-title":"Life cycle assessment of artificial intelligence applications: research gaps and opportunities","volume":"135","author":"Plociennik","journal-title":"Procedia CIRP"},{"issue":"11","key":"R10","doi-asserted-by":"publisher","first-page":"2390","DOI":"10.1007\/s11367-024-02419-2","article-title":"A critical analysis of global warming potential of data centers in the digital era","volume":"30","author":"Bux","journal-title":"Int J Life Cycle Assess"},{"issue":"6","key":"R11","doi-asserted-by":"publisher","first-page":"064017","DOI":"10.1088\/1748-9326\/abfba1","article-title":"The environmental footprint of data centers in the United States","volume":"16","author":"Siddik","journal-title":"Environ Res Lett"},{"key":"R12","unstructured":"AI Environmental Footprint Calculator. Kaveh Mozafari. URL: https:\/\/www.kavehmozafari.com\/calc [Accessed 30-06-2026]"},{"key":"R13","unstructured":"Anthony LFW Kanding B Selvan R . Carbontracker: tracking and predicting the carbon footprint of training deep learning models. arXiv. Preprint posted online on Jul 6, 2020. [doi: 10.48550\/arXiv.2007.03051]"},{"key":"R14","doi-asserted-by":"crossref","unstructured":"Lacoste A Luccioni A Schmidt V Dandres T . Quantifying the carbon emissions of machine learning. arXiv. Preprint posted online on Oct 21, 2019. [doi: 10.48550\/arXiv.1910.09700]","DOI":"10.5260\/chara.21.2.8"},{"key":"R15","volume":"21","author":"Henderson","journal-title":"J Mach Learn Res"},{"issue":"12","key":"R16","doi-asserted-by":"publisher","DOI":"10.1002\/advs.202100707","article-title":"Green algorithms: quantifying the carbon footprint of computation","volume":"8","author":"Lannelongue","journal-title":"Adv Sci (Weinh)"}],"container-title":["JMIR AI"],"original-title":[],"language":"en","deposited":{"date-parts":[[2026,8,18]],"date-time":"2026-08-18T19:44:38Z","timestamp":1787082278000},"score":1,"resource":{"primary":{"URL":"https:\/\/ai.jmir.org\/2026\/1\/e90770"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,18]]},"references-count":16,"URL":"https:\/\/doi.org\/10.2196\/90770","relation":{},"ISSN":["2817-1705"],"issn-type":[{"value":"2817-1705","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,18]]},"article-number":"v5i8062e90770"}}