From Efficiency Gains to Rebound Effects: The Problem of Jevons' Paradox in AI's Polarized Environmental Debate
2025/01/27 by Alexandra Sasha Luccioni, Emma Strubell, Luccioni, Alexandra Sasha +3 · 14 voices · 17 citations
Energy · Environmental Science · Economics, Econometrics and Finance · #Energy, Environment, and Transportation Policies #Environmental Impact and Sustainability #Climate Change Policy and Economics
paper · pdf · doi:10.1145/3715275.3732007
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-based dialogue on AI's role in the climate crisis.
Citations
Cited by
Discussions
- New @sashamtl.bsky.social @strubell.bsky.social @katecrawford.bsky.social paper is a v v good read -->>> "many of the purported “solutions” to climate challenges via AI remain tethered to profit-driv [bsky, 53 points, 1 comments]
- As AI systems grow more efficient, will they reduce environmental impacts —or make them worse? Our new paper with @strubell.bsky.social and Kate Crawford explores how 𝗝𝗲𝘃𝗼𝗻𝘀’ 𝗣𝗮𝗿𝗮𝗱𝗼𝘅 app [bsky, 36 points, 1 comments]
- Le paradoxe de Jevons (qui ne s'est jamais vraiment démenti): les gains d'efficacité dans les technologies mènent à une plus grande consommation de ressources si ces derniers mènent à une multiplicati [bsky, 21 points, 3 comments]
- Want to read the whole paper? 🌍 "From Efficiency Gains to Rebound Effects: The Problem of Jevons’ Paradox in AI’s Polarized Environmental Debate" Pre-print here: arxiv.org/pdf/2501.16548 [bsky, 20 points, 0 comments]
- Another thing: Matter and Space's video ends with a call for tech companies to prioritize efficiency to mitigate AI's environmental harms. However, as Sasha Luccioni, Emma Strubell, and Kate Crawford [bsky, 6 points, 0 comments]
- Excellent paper from Dr Sasha Luccioni on the paradoxical conversations about AI and the environment arxiv.org/pdf/2501.16548 #ArtificialIntelligence #AIEducation #AIEdu #AIInEd #AIInEdu #EduSky [bsky, 4 points, 0 comments]
- ...from the growing energy demands and corresponding negative environmental impacts of AI. Check out our new preprint discussing what *actually* matters in the "debate" about AI, energy and the envi [bsky, 3 points, 1 comments]
- amg vc pediu em um outro tuite artigos sobre IA, li esse hoje sobre os impactos diretos e indiretos, achei bem abrangente arxiv.org/pdf/2501.16548 [bsky, 2 points, 1 comments]
- arxiv.org/abs/2501.165... [bsky, 1 points, 0 comments]
- Great new paper (pre-print) by Sasha Luccioni et al. on the many issues with sustainability claims of generative AI: "This paper examines how the problem of Jevons’ Paradox applies to AI, whereby ef [bsky, 1 points, 1 comments]
- C'est un peu hors sujet, mais il y a un essai sur l'effet rebond lié aux récents progrès en efficacité des modèles d'IA génératives qui est discuté par ici, et qui pourrait te plaire! arxiv.org/pdf/25 [bsky, 1 points, 1 comments]
- From Efficiency Gains to Rebound Effects: The Problem of Jevons’ Paradox in AI’s Polarized Environmental Debate PAPER: arxiv.org/pdf/2501.16548 #sustainableit #greenit #ai #sustainability [bsky, 0 points, 0 comments]
- It's a preprint (and full disclosure I haven't read it in full yet) but is this maybe the kind of thing you're after? arxiv.org/pdf/2501.16548 [bsky, 0 points, 1 comments]
- Did you know AI’s climate impact is more complex than we think? A recent paper on arXiv dives into Jevons’ Paradox—the idea that AI’s efficiency gains can backfire, driving more resource use, not less [bsky, 0 points, 0 comments]
Related