From Efficiency Gains to Rebound Effects: The Problem of Jevons' Paradox in AI's Polarized Environmental Debate
Open accessThe DOI matches its Crossref record; 105 of the 106 references it cites were found in a registry or an index, and 1 are not checked yet; 39 of the 103 sources its claims cite were checked for support.
Abstract
As the climate crisis deepens, artificial intelligence (AI) has emerged as a contested force: some champion its potential to advance renewable energy, materials discovery, and large-scale emissions monitoring, while others underscore its growing carbon footprint, water consumption, and material resource demands. Much of this debate has concentrated on direct impacts—energy and water usage in data centers, e-waste from frequent hardware upgrades—without addressing the significant indirect effects. This paper examines how the problem of Jevons’ Paradox applies to AI, whereby efficiency gains may paradoxically spur increased consumption. We argue that understanding these second-order impacts requires an interdisciplinary approach, combining lifecycle assessments with socio-economic analyses. Rebound effects undermine the assumption that improved technical efficiency alone will ensure net reductions in environmental harm. Instead, the trajectory of AI’s impact also hinges on business incentives and market logics, governance and policymaking, and broader social and cultural norms. We contend that a narrow focus on direct emissions misrepresents AI’s true climate footprint, limiting the scope for meaningful interventions. We conclude with recommendations that address rebound effects and challenge the market-driven imperatives fueling uncontrolled AI growth. By broadening the analysis to include both direct and indirect consequences, we aim to inform a more comprehensive, evidence-b…
Verified 2026-11-22 · 1 reference not checked yet
Overview
CoChat Summary
TL;DR
This conceptual paper argues that the polarized debate over AI's environmental impact is stuck on direct effects (data center energy, water, e-waste) while ignoring indirect "rebound effects," in which efficiency gains spur higher overall consumption—Jevons' Paradox. The authors call for interdisciplinary lifecycle-plus-socioeconomic analysis and for transparency, regulation, and business-model reform rather than efficiency alone.
Key findings
- Direct impacts are rising despite efficiency gains: the IEA puts data center electricity at 2% of global electricity, more than doubling by 2026 (surpassing Canada's national power use); Google reports a 48% GHG increase since 2019, Baidu 32.6% over 2021, and Microsoft 29.1% since 2020, while one analysis found real data center footprints can exceed reported values by over 600%.
- Water and materials burdens are concentrated and opaque: Microsoft reported a 34% increase in global water consumption between 2021 and 2022 (over 1.7 billion gallons) and Google a 20% uptick; one study suggests 10–50 GPT-3 queries consume about half a liter of water; 54% of technology-critical materials come from Indigenous or peasant territory and 62% from drought-prone zones.
- E-waste and hardware demand scale with AI: e-waste reached 62 million tonnes in 2022 with only about 22% formally collected and recycled, and AI is estimated to add 1.2–5 million metric tons of e-waste by 2030; NVIDIA shipped 3.7 million GPUs in 2024, more than a million more units than in 2023, even as hardware efficiency improved.
- The paper taxonomizes indirect effects into material (substitution, scale, space), economic (direct, indirect/real income, economy-wide), and societal/behavioral (induction, time, indirect policy) rebound effects—e.g., targeted advertising revenue reached $36 billion in 2024 and Amazon's AI recommendation tool generates almost a third of its annual sales, while AI-assisted navigation has been shown to cut emissions by an average of 3.4%.
Methods
- Conceptual/position paper: no new empirical data collection or experiment; it synthesizes existing literature, corporate ESG reports, and prior studies to reframe the AI-and-environment debate around indirect effects.
- Builds an analytical taxonomy by adapting Börjesson Rivera et al.'s second-order ICT effects taxonomy and Kaack et al.'s work on AI's indirect impacts, organizing rebound effects into three themes: material objects and physical spaces, micro-/macro-economic processes, and society and human behavior.
- Proposes mitigation and tracking approaches (qualitative/quantitative research, transparency, regulation, efficient devices and distributed compute) rather than testing them.
Limitations
- Stated: the authors note that applying Jevons' Paradox to AI has "notable conceptual and empirical limitations," citing Sorrell's critique that it oversimplifies causal relationships among efficiency gains, economic expansion, and behavioral responses; they also state that rebound effects are characterized by uncertainty and heterogeneity, and that measuring them is hampered by severe data scarcity, trade-secret opacity, and non-standardized corporate reporting.
- Stated: the paper acknowledges that claims about AI's climate-positive applications "remain largely hypothetical, with little explicit evidence or quantitative analysis," and that existing AI-vs-human LCA comparisons (e.g., 130–1500× less CO2e per page of text, 310–2900× per image) do not account for social impacts, legality, or rebound effects.
- Evident from the text: the paper offers no original quantification of rebound magnitudes for AI and does not empirically validate its taxonomy; several cited figures come from corporate self-reports or third-party estimates whose reliability the authors themselves question.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
28 Valid · 8 Not valid · 44 Not checked · 94 Own results and interpretations
At least one cited source doesn’t support this claim or points to a different paper, and none confirmed it — see the verdicts below.
In the paper
“a life cycle assessment (LCA), which evaluates the environmental impacts of an artifact arising throughout its existence (typically including disposal), has been performed comparing print books to e-readers, finding that 115 books would produce the same amount of CO2 as a single Amazon Kindle device [32, 103].”
What the sources say
- Disputedcites [32] [32] Dealva Jade Dowd-Hinkle. 2012. Kindle vs. Printed book an environmental analysis . Rochester Institute of Technology. -Experimental — accuracy not yet measured
The paper does compare print books and e-readers using life-cycle assessment, but it does not report a 115-book CO2 equivalence. Section 2.5, “Print vs. Screen” (page not supplied), gives estimates of more than 23 books or about 4 books to offset a Kindle’s emissions, depending on the analysis. (LLM, conf=0.88)
Checked by a reasoning model — may be wrong.
- Not checked yetcites [103] [103] Joy Scrogum. 2009. Books vs. eBooks -A life cycle comparison. https://sustainable-electronics.istc.illinois.edu/2009/11/05/books-vsebooks-a-life-cycle-comparison/ -
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Disputed
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“At a more granular level, hardware optimizations such as parallelization, do in fact support these economies of scale; e.g. batched inference on a GPU scales sub-linearly in the number of examples, allowing multiple queries to be performed at once at only a marginally higher cost [109].”
What the sources say
- Partly supportedcites [109] [109] Franyell Silfa, Jose Maria Arnau, and Antonio González. 2022. E-BATCH: Energyefficient and high-throughput RNN batching. ACM Transactions on Architecture and Code Optimization (TACO) 19, 1 (2022), 1-23. - sourceExperimental — accuracy not yet measured
Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“as Koomey and Masanet [65] cogently argue, this is not the first time the alarm has been raised about rising energy use due to technology, and that similar projections made in the dot-com boom of the early 2000s failed to materialize.”
What the sources say
- Partly supportedcites [65] [65] Jonathan Koomey and Eric Masanet. 2021. Does not compute: Avoiding pitfalls assessing the Internet's energy and carbon impacts. Joule 5, 7 (2021), 1625-1628. https://doi.org/10.1016/j.joule.2021.05.007 - sourceExperimental — accuracy not yet measured
The Introduction (page unavailable) says extrapolations of 1990s Internet-traffic growth led to excess network capacity around 2000, with significant energy implications. It also describes later projections of data-center energy use that overestimated growth. It does not say that projections of rising energy use made during the dot-com boom failed to materialize. (LLM, conf=0.86)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“one recent study estimates that AI will generate an additional 1.2-5 million metric tons of e-waste by 2030 [134].”
What the sources say
- Partly supportedcites [134] [134] Peng Wang, Ling-Yu Zhang, Asaf Tzachor, and Wei-Qiang Chen. 2024. E-waste challenges of generative artificial intelligence. Nature Computational Science (2024), 1-6. - sourceExperimental — accuracy not yet measured
The paper estimates substantial LLM-related server e-waste through 2030, so it supports the general point. But the results section gives cumulative estimates of 8 million tons in its most conservative scenario and about 16 million tons in its optimistic scenario, not the cited 1.2–5 million-ton range. The cited sentence does not specify a different measure, such as waste generated in 2030 alone. (LLM, conf=0.84)
hedgedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“This, in turn, has impacts on supply chain dynamics, redistributing the environmental footprint of operations (e.g. heating and cooling), logistics and transportation [41].”
What the sources say
- Partly supportedcites [41] [41] Bastien Girod, Peter de Haan, and Roland W Scholz. 2011. Consumption-asusual instead of ceteris paribus assumption for demand: Integration of potential rebound effects into LCA. The International Journal of Life Cycle Assessment 16 (2011), 3-11. - sourceExperimental — accuracy not yet measured
The Abstract says product changes can lead households to reallocate consumption, changing environmental impact. But the paper does not link e-commerce or changes in physical space to supply-chain operations, heating and cooling, logistics, or transportation footprints; its mention of online shopping in §4.2 concerns consumer preferences. (LLM, conf=0.78)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
70 more claims · 11 with their cited sources checked
Create a free account to see all 174 claims and to verify the 31 references still open.
- It may state two separate claims as one. 47
- Probably not a substantial claim. 73
- The quote only partly states this claim. 1
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 2.1 Arguments that AI is a net climate negative · p. 2
“Data from the International Energy Agency (IEA) has provided the strongest international benchmark.”
§ References · p. 12
“- [56] International Energy Authority. 2023. Data Centres and Data Transmission Networks.”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 3.2 Economic Rebound Effects · p. 6
“NVIDIA shipped 3.7 million GPUs in 2024 (more than a million more units than in 2023) due to increased demand, despite these improvements in efficiency [105].”
§ References · p. 13
“- [105] Agam Shah. 2024. Nvidia Shipped 3.76 Million Data-center GPUs in 2023, According to Study.”
Citations
Cites
DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
DeepSeek-AI, Guo, D., Yang, D., Zhang, H., Song, J., Wang, P., Zhu, Q., Xu, R., Zhang, R., Ma, S., et al. (2025) · arXiv · doi:10.48550/arxiv.2501.12948 · Cited by 18,918
Scaling Laws for Neural Language Models
Kaplan, J., McCandlish, S., Henighan, T., Brown, T. B., Chess, B., Child, R., Gray, S., Radford, A., Wu, J., & Amodei, D. (2020) · arXiv · doi:10.48550/arxiv.2001.08361 · Cited by 9,550
In-Datacenter Performance Analysis of a Tensor Processing Unit
Jouppi, N. P., Young, C., Patil, N., Patterson, D., Agrawal, G., Bajwa, R., Bates, S., Bhatia, S., Boden, N., Borchers, A., et al. (2017) · Proceedings of the 44th Annual International Symposium on Computer Architecture · doi:10.1145/3079856.3080246 · Cited by 4,537
Energy and Policy Considerations for Deep Learning in NLP
Strubell, E., Ganesh, A., & McCallum, A. (2019) · arXiv · doi:10.48550/arxiv.1906.02243 · Cited by 3,735
Self-citationEnergy and Policy Considerations for Deep Learning in NLP
Strubell, E., Ganesh, A., & McCallum, A. (2019) · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/p19-1355 · Cited by 3,684 · cited 2×
“On the one hand, it is champi-oned as a driver of efficiency that could potentially invent new solutions to the climate crisis or reinvent old ones [5, 21, 99]; on the other hand, it is criticized for its steeply increasing resource demands and carbon footprint [68, 120, 134].”
Self-citationThe role of artificial intelligence in achieving the Sustainable Development Goals
Vinuesa, R., Azizpour, H., Leite, I., Balaam, M., Dignum, V., Domisch, S., Felländer, A., Langhans, S. D., Tegmark, M., & Fuso Nerini, F. (2020) · Nature Communications · doi:10.1038/s41467-019-14108-y · Cited by 3,294 · cited 1×
“There are several key areas in which this is particularly salient, from generating insights from large quantities of data in different modalities to optimizing complex systems and tools [21].”
Reassessing the projections of the World Water Development Report
Boretti, A., & Rosa, L. (2019) · npj Clean Water · doi:10.1038/s41545-019-0039-9 · Cited by 2,554 · cited 1×
“By 2050, 50% of the world’s population is projected to live in a region affected by water scarcity [14], but the full impacts of data center water usage is unknown as “the entire data centre industry suffers from a lack of transparency” [80].”
GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models
Eloundou, T., Manning, S., Mishkin, P., & Rock, D. (2023) · arXiv · doi:10.48550/arxiv.2303.10130 · Cited by 2,330
The environmental price of fast fashion
Niinimäki, K., Peters, G., Dahlbo, H., Perry, P., Rissanen, T., & Gwilt, A. (2020) · Nature Reviews Earth & Environment · doi:10.1038/s43017-020-0039-9 · Cited by 2,034 · cited 1×
“However, the amount of plastics, petrochemicals and waste that are engendered by the increased accessibility of these items and therefore their increased consumption are damaging to the environment [81].”
E-waste: An assessment of global production and environmental impacts
Robinson, B. H. (2009) · Science of The Total Environment · doi:10.1016/j.scitotenv.2009.09.044 · Cited by 1,777 · cited 1×
“The UN’s Global E-Waste Monitor 2024 showed that about 22% of e-waste has been shown tobe formally collectedand recycled, with global generation of electronic waste rising five times faster than e-waste recycling [10].”
Research works printed but not linked: 9
- [12] Yoshua Bengio, Salem Lahlou, Tristan Deleu, Edward J Hu, Mo Tiwari, and Emmanuel Bengio. 2023. Gflownet foundations. The Journal of Machine Learning Research 24, 1 (2023), 10006-10060. -
- [19] Ye Chen, Michael Kapralov, John Canny, and Dmitry Pavlov. 2009. Factor modeling for advertisement targeting. Advances in neural information processing systems 22 (2009). -
- [21] Peter Clutton-Brock, David Rolnick, Priya L Donti, and Lynn Kaack. 2021. Climate change and AI. recommendations for government action . Technical Report. GPAI, Climate Change AI, Centre for AI & Climate. -
- [39] Jaume Freire-González and Ignasi Puig-Ventosa. 2015. Energy efficiency policies and the Jevons paradox. International Journal of Energy Economics and Policy 5, 1 (2015), 69-79. -
- [60] WStanley Jevons. 1866. The coal question. In The Economics of Population . Routledge, 193-204. -
- [75] Eric Masanet, Arman Shehabi, Nuoa Lei, Sarah Smith, and Jonathan Koomey. 2020. Recalibrating global data center energy-use estimates. Science 367, 6481 (2020), 984-986. - · DOI · Paper page
- [121] Richard Sutton. 2019. The bitter lesson. Incomplete Ideas (blog) 13, 1 (2019), 38. -
- [123] Zeeshan Tariq, Murtada Saleh Aljawad, Amjed Hasan, Mobeen Murtaza, Emad Mohammed, Ammar El-Husseiny, Sulaiman A Alarifi, Mohamed Mahmoud, and Abdulazeez Abdulraheem. 2021. A systematic review of data science and machine learning applications to the oil and gas industry. Journal of Petroleum Exploration and Production Technology (2021), 1-36. -
- [127] Sunita Tiwari and Pratibha Singh. 2011. Environmental impacts of e-commerce. In International conference on environment Science and engineering , Vol. 8. 202207. -
Cited by
Two types of AI existential risk: decisive and accumulative
Kasirzadeh, A. (2025) · Philosophical Studies · doi:10.1007/s11098-025-02301-3 · Cited by 115
“Although no single AI application may pose an immediate existential threat, the aggregated impact of AI across various adapted subsystems could lead to critical systemic imbalances or crises, when reaching a tipping point.”
Inie, N., Falk, J., & Selvan, R. (2025) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3706598.3714227 · Cited by 43
“A second discrepancy revolves around the unintended effects increasing efficiency might have, such as when a reduction in compute leads greater overall use of that system over time [211, 342].”
Cugurullo, F., Caprotti, F., Day, J., Geoghegan, S., Lynch, C. R., Menga, F., Robinson, C., & Williams, J. (2025) · Urban Political Ecology · doi:10.1177/30497515251344495 · Cited by 31
“Certainly, the rapidity of the emergence of AI has become a major challenge for urban energy transitions because its applications, such as AVs, robots and drones, IoT, software and apps, as well as the data processing (Cugurullo et al., 2023; Luccioni et al., 2025), require energy.”
Electrification, flexibility or both? Emerging trends in European energy policy
Mellot, A., Moretti, C., Nuñez-Jimenez, A., Lindér, J., Moro, N., Powell, S., Markard, J., Winzer, C., & Patt, A. (2025) · Energy Policy · doi:10.1016/j.enpol.2025.114725 · Cited by 17
Chouksey, A., Rajan, A. K., Gurjar, V., Tiwari, R., & Mishra, P. K. (2026) · Global Environmental Change Advances · doi:10.1016/j.gecadv.2025.100029 · Cited by 16
Smart sensors and AI: Sustainable solutions for environmental pollution challenges
Mutlag, F., Elaibi, H. K., Mahious, R., Halvaci, E., & Şen, F. (2025) · Next research. · doi:10.1016/j.nexres.2025.101195 · Cited by 15
Kim, M., Yoo, T.-A., & Chung, J. (2026) · IEEE Access · doi:10.1109/access.2026.3659894 · Cited by 24
“…consumption and greenhouse The associate editor coordinating the review of this manuscript and approving it for publication was Mu-Yen Chen . gas (GHG) have intensified, fueling ongoing debates about whether AI serves as a net positive or a net negative for the environment [1], [2], [3], [4].”
A Guide to Machine Learning Epistemic Ignorance, Hidden Paradoxes, and Other Tensions
Naser, M. Z. (2025) · Wiley Interdisciplinary Reviews Data Mining and Knowledge Discovery · doi:10.1002/widm.70038 · Cited by 17
“• The compute-efficiency paradox (Sasha Luccioni et al. 2025): Model compression, pruning, and quantization free up hardware cycles, yet the saved compute is quickly rein-vested to train even larger models, driving aggregate energy use upward.”
The role of mineral raw material imports in driving the energy transition
Fikru, M. G., & Kilinc‐Ata, N. (2025) · Mineral Economics · doi:10.1007/s13563-025-00570-2 · Cited by 8
Davis, J. L., & Sloane, M. (2025) · Social Science Computer Review · doi:10.1177/08944393251394690 · Cited by 8
“The time to mobilize this latter power is now, as AI ’ s peak coincides with, and is implicated in, the production and ampli fi cation of structural inequities (Davis et al., 2021; Joyce & Cruz, 2024; Joyce et al., 2021; Sloane & Moss, 2019) democratic erosion (Coeckelbergh, 2024; Kreps & Kriner, 2023), geopolitical co…”
Ground-truthing AI energy consumption: validating CodeCarbon against external measurements
Fischer, R. (2026) · it - Information Technology · doi:10.1515/itit-2025-0031 · Cited by 36
“Given that researchers often fail to justify their compute expenses or discuss potential negative consequences [4], [5], the observable “bigger-is-better paradigm” [6] is highly concerning but not surprising.”
From computation to environmental cost the resource burden of artificial intelligence
Falk, S., Corrêa, N. K., Luccioni, A. S., Biber‐Freudenberger, L., & van Wynsberghe, A. (2026) · Communications Earth & Environment · doi:10.1038/s43247-026-03537-5 · Cited by 28
Self-citationThe unseen carbon cost of AI workforce: A behavioral theory perspective of environmental scalability
Lv, D. D., & Cho, E. (2025) · Business Horizons · doi:10.1016/j.bushor.2025.07.005 · Cited by 8
AI, Digital Platforms, and the New Systemic Risk
Hacker, P., Edwards, L., & Kasirzadeh, A. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3806506 · Cited by 19
“AI has significant potential for reducing emissions [86, 95], but the current tendency is for AI to generate more demand for energy [50, 55, 70, 72].”
Quantifying the sustainability impacts of smart cities and AI: A meta-analysis
Hwang, M. I., & Verhulsdonck, G. (2026) · Sustainable Cities and Society Advances · doi:10.1016/j.scsadv.2026.100027 · Cited by 10
Uncertain quantum computing futures and potential energy and physical resource impacts at scale
McCollum, D. L., Sharma, N., Pan, M., Yao, W., Bhanja, S. N., Chehade, S., & Delgado, A. (2026) · Renewable and Sustainable Energy Transition · doi:10.1016/j.rset.2026.100140 · Cited by 5
The Energy and Environmental Footprint of AI
Vandenbergh, M. P., Thorpe, E. I., & Gilligan, J. M. (2026) · Michigan Journal of Environmental & Administrative Law · doi:10.36640/mjeal.15.2.energy · Cited by 2
Zhou, M., & Javed, M. A. (2026) · Geological Journal · doi:10.1002/gj.70174 · Cited by 2
Post-growth Perspectives in HRI
Thunberg, S., Gamboa, M., Torre, I., & Penzenstadler, B. (2026) · ACM/IEEE International Conference on Human-Robot Interaction (HRI) · doi:10.1145/3757279.3788811 · Cited by 2
Statactivism Up and Down the Stack: Dilemmas in the Estimation of AI’s Environmental Footprint
Lenoir, T., & Parker, C. (2025) · Digital Society · doi:10.1007/s44206-025-00196-5 · Cited by 4
“In a recent effort to raise attention on the full range of effects of AI on the environment, Luccioni et al. ( 2025) have once again called upon the consideration of so-called rebound effects in global quantification efforts.”
Resource-Aware Machine Learning
Morik, K. J. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-03028-3_7 · Cited by 1
The impact of new infrastructure construction on green high-quality development in China
Sun, Y., & Masron, T. A. (2025) · Discover Sustainability · doi:10.1007/s43621-025-02149-x · Cited by 3
“This has spawned high-energy-consuming applications like generative artificial intelligence, which were previously unfeasible due to excessive costs, resulting in hyperelastic demand growth [74, 85].”
Vrain, E., Bieser, J. C. T., & Wilson, C. (2025) · Energy and Buildings · doi:10.1016/j.enbuild.2025.116920 · Cited by 3
Ensuring Sustainable Digital Marketing Through the Integration of ESG Principles
Thangam, D., Kumar, G. P., Arumugam, T., et al. (2026) · World sustainability series · doi:10.1007/978-3-032-01715-4_8 · Cited by 1
The iron triangle in the age of AI?
Bowker, L. (2026) · Digital Translation · doi:10.1075/dt.00016.bow · Cited by 1
Designing Integrated Policies for the Twin Transition: Challenges and Tradeoffs
Maggetti, M. (2026) · Policy Studies Journal · doi:10.1111/psj.70118 · Cited by 2
Thompson-Olson, M., Michael, D., & Saluja, J. (2026) · American Journal of Lifestyle Medicine · doi:10.1177/15598276261449784 · Cited by 2
Regulating AI: Where U.S. State Policy and HCI (Mis)align
Migineishvili, N., Gao, A., Adjagbodjou, A. D., Thakur, D., Just, R., & Reinecke, K. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791424 · Cited by 1
“The enormous energy demands of training large-scale models [24, 53, 102], for instance, were rarely documented, even as such demands threaten to undermine already fragile climate commitments across both government and industry [84, 109].”
Balancing AI’s energy consumption with commitments to clean energy (SDG 7): Is there a trade-off?
Glavas, D. (2025) · International Journal of Green Energy · doi:10.1080/15435075.2025.2546047 · Cited by 2
“…the notion of resource prioritization in sustainability strategies (Srivastava et al. 2023), and concerns about trade-offs in operational implementation (Luccioni, Strubell, and Crawford ,2025), we use two proxies to evaluate energy-efficient AI practices: monitoring and overengineering avoidance.”
InfluentialOpenAI’s newly launched Sora 2 makes AI’s environmental impact impossible to ignore
Diab, R. (2025) · The Conversation · doi:10.64628/aam.c3aknrke4
The Double-Edged Sword: Evaluating the Climate Impact of Speech Processing Technologies
Saber, A., Elbedwehy, S., Hassan, E., & Emara, T. Z. (2025) · IGI Global eBooks · doi:10.4018/979-8-3373-3048-8.ch010
Global Value Chains, Digital Tech, Systemic EcocidePreprint
Selwyn, B. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6036078
Semiconductors, AI, and Sustainability: A Convergent Pathway for Clean Energy and Smart Cities
Orduño-Osuna, J. H., Mercado-Herrera, A., & Ling-López, J. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-3481-3.ch011 · Cited by 1
FlipFlop: A Static Analysis-based Energy Optimization Framework for GPU Kernels
Rajput, S., Brandt, A., Elisseev, V., & Sharma, T. (2026) · ACM/IEEE International Conference on Software Engineering: Companion Proceedings (ICSE-Companion) · doi:10.1145/3744916.3787793 · Cited by 4
The Climate Impact Paradox of Artificial Intelligence: Literature Review and Research Prospects
Chen, Z., Xiong, Q., Hu, Z., Liu, S., & Ma, J. (2026) · Journal of Energy and Climate Change · doi:10.3724/j.issn.2097-4981.jecc-2025-0385
Energy and physical resource impacts of quantum computing merit greater attention
McCollum, D. L., Bhanja, S. N., Chehade, S., Delgado, A., Pan, M., & Sharma, N. (2026) · Nature Reviews Clean Technology · doi:10.1038/s44359-026-00156-3 · Cited by 2
“While quantum technology could promise advancements in compute efficiency, those efficiency gains could simultaneously lead to increases in demands for new services, including heretofore unimagined computing possibilities 11 .”
AI for climate: from technological solutions to relational accountability
Stein, S. (2026) · npj Climate Action · doi:10.1038/s44168-026-00376-0
The Footprint of the Algorithm
Fernandes, C. E., Morais, R., & Piñeiro-Naval, V. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-7200-6.ch011
Milou, A. C. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6679558
Lupton, D., & Bailey‐Charteris, B. (2026) · Environmental Communication · doi:10.1080/17524032.2026.2673346 · Cited by 2
The Jevons Paradox in AI Infrastructure: Why Efficiency Gains Accelerate Compute ConsumptionPreprint
Do, T. A. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6775299
Fernandes, C. E., Morais, R., & Piñeiro-Naval, V. (2026) · Metrics · doi:10.3390/metrics3020009
“In contrast, others already point to the harms and dangers posed by overconsumption and blind trust in AI [15–17].”
Greenwashing AI in Higher Ed: An environmental argument for universities’ resistance to Big Tech
Riemland, M. (2026) · Open Research Europe · doi:10.12688/openreseurope.23757.1
Kumar, M., Patouillard, L., Donaldson, A., Cornago, S., Bartocci, P., Tu, Q., Wachter, H. D., Hübner, N., Dumont, M., Coroamă, V. C., et al. (2026) · The International Journal of Life Cycle Assessment · doi:10.21203/rs.3.rs-9984806/v1
Petmesidou, M. (2026) · Global Social Policy · doi:10.1177/14680181261456072
Public perceptions of AI energy consumption: Awareness, responsibility attribution, and strategies
Klesel, M., & Messer, U. (2026) · Electronic Markets · doi:10.1007/s12525-026-00908-w · Cited by 2
“The environmental impact of GenAI use is clearly a many-hands problem, and the impact of a single user’s actions is not straightforward to quantify, yet the cumulative actions of all users have a huge effect (Luccioni et al., 2024b, 2025).”
Abidemi, S. U., Yanaşoğlu, M. M., Mlimwa, C. P., Alturjman, S., Al-Turjman, F. M., & Mohapatra, H. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-2600-2761-5.ch008
Mapping Emerging Climate Misinformation Playbooks in the Global South
Locatelli, M. S., Dong, W., Alzamora, P. L., Dutenhefner, P. R., Meira, W., Cha, M., & Almeida, V. A. F. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3806492 · Cited by 1
de Walle, A. V., & Aerts, G. (2026) · Business Strategy and the Environment · doi:10.1002/bse.71232
Billah, W., Suarez, J., & Wolters, E. A. (2026) · Journal of Science Policy & Governance · doi:10.38126/jspg280101
Empowerment or constraint? The impact of digital intelligence investment on energy efficiency
Zhang, C., Hu, W., Liu, Y., et al. (2026) · Applied Economics Letters · doi:10.1080/13504851.2026.2701286
Akpan, M. N. U., & Adebayo, A. O. (2026) · Reporting and Accountability Review · doi:10.22495/rarv3i1p3 · Cited by 1
Global value chains, digital tech, systemic ecocide
Selwyn, B. (2026) · Globalizations · doi:10.1080/14747731.2026.2709957
KI-Lernen: das kostet Energie … !?
Friederichs, H. (2026) · doi:10.59350/jgh8f-c8q05
Li, Z., Wang, M., Zhou, T., & Zhang, C. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7244749
KI-Lernen: das kostet Energie … !?
Friederichs, H. (2026) · doi:10.59350/h9j9p-kcn04
No Exchange Rate: Carbon in the Economics of AI ComputePreprint
Murphy, J. E. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7216759
When Innovation Outpaces Sustainability
Poovayar Priya, M. (2026) · IGI Global eBooks · doi:10.4018/979-8-3373-8490-0.ch011
AI-Enabled Business Models for Sustainability: Pathways to Regenerative Entrepreneurial Growth
Ramachandran, R. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-7699-8.ch003
Palicki, S.-K., Bravo, I., & Claessen, C. (2026) · Social Science Computer Review · doi:10.1177/08944393261473230
Who Owns the Wai (Water)? AI, Power, and the Public Trust
Hester, J. D. (2026) · Communications of the ACM · doi:10.1145/3785371
The CO2ST of Agentic AI in Research
Falk, J., Inie, N., & Selvan, R. (2026) · Nordic Conference on Human-Computer Interaction · doi:10.1145/3821402.3830115
Metadata
- Authors
- Luccioni, A. S., Strubell, E., & Crawford, K.
- Affiliations
- Hugging Face, Montreal, Canada
- Carnegie Mellon University (US)
- Microsoft Research New York City (United States) (US)
- Venue
- Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency
- Publisher
- ACM
- Published
- 2025-06-23
- Pages
- 76-88
- OpenAlex
- W4411541921
- Semantic Scholar
- fff2147b809fd64cb79bc0b8f3c762dc2c7e6787
- Cited by
- 71
- References
- 87
- Citations by year
- 2026: 53 · 2025: 18
- FWCI
- 150.99
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Locations
- ACM Conference on Fairness, Accountability, and Transparency (FAccT) (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion, cc-by-nc-sa)
- arXiv (Cornell University)
- Funders
- NSF (National Science Foundation) (2326610)
- Topics
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Green IT and Sustainability
- Physical Sciences › Environmental Science › Environmental Engineering › Environmental Impact and Sustainability
- Social Sciences › Economics, Econometrics and Finance › Economics and Econometrics › Energy, Environment, Economic Growth
- Keywords
- Jevons paradox, rebound effect, artificial intelligence, environmental impact of AI, efficiency improvement, energy consumption, environmental debate, technology and sustainability
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-27, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-27 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.