Cross-corpus research synthesis
Environmental Impact of AI Training
Energy consumption, water usage, carbon emissions, and resource demands of large-model training + inference. EU AIA Recital 142 + Art. 95 voluntary codes; NIST AI 600-1 Environmental Impacts (named risk category); G7 Hiroshima Code §6 sustainable AI; emerging French ARCEP + Spanish AI Bill obligations; SDG-linked references in UN + AU + ASEAN frameworks.
Synthesised deterministically from 13 articles that engage this theme. Empirical consensus: emerging · contested: Should environmental obligations attach to (a) model-provider disclosure, (b) datacenter operator emissions caps, or (c) end-customer reporting? The training-vs-inference split also remains unresolved across instruments.. Full theme article: /wiki/environmental-impact-of-training. Machine-readable: /wiki/synthesis.json.
Cross-jurisdiction stances (2 govern, 12 engage)
| Instrument | Verdict | Provision excerpt / citation |
|---|---|---|
| EU AI Act | implicit | Codes of conduct may cover … assessing and minimising the impact of AI systems on environmental sustainability, including as regards energy-efficient programming and techniques for design… (paraphrase) Art. 95 voluntary codes of conduct include environmental sustainability; Recital 142 references energy efficiency reporting for GPAI |
| Executive Order 14110 on Safe, Secure, Trustworthy AI | implicit | §5.2 directs environmental-review consideration; §4.2 reporting includes some energy data |
| G7 Hiroshima AI Process Code of Conduct | implicit | Code §6 references sustainable AI development; not detailed obligation |
| OECD AI Principles (Recommendation) | implicit | Principle 1.1 inclusive growth + sustainable development; addresses environment implicitly |
| Council of Europe Framework Convention on AI | implicit | Art. 7 sustainability principle; environmental impact subsumed |
| UN GA Resolution on Safe, Secure, Trustworthy AI | implicit | Preamble references SDGs which include climate goals |
| NIST AI RMF Generative AI Profile | governs | NIST AI 600-1 — Environmental Impacts is one of 12 named GenAI risk categories |
| ASEAN Guide on AI Governance and Ethics | implicit | Guide references sustainable AI principles; not operationalised |
| African Union Continental AI Strategy | implicit | Continental strategy includes sustainability themes; not operationalised |
| UNESCO Recommendation on the Ethics of Artificial Intelligence | governs | “Member States and business enterprises should assess the direct and indirect environmental impact throughout the AI system life cycle, including... its carbon footprint, energy consumption” Policy Area 'Environment and Ecosystems', para 84 — assess direct/indirect environmental impact incl. carbon footprint + energy consumption |
| Italy Law No. 132/2025 on Artificial Intelligence (Legge 23 settembre 2025, n. 132) | implicit | “… nel rispetto … dei princìpi di trasparenza, proporzionalità, sicurezza, protezione dei dati personali, riservatezza, accuratezza, non discriminazione, parità dei sessi e sostenibilità.” Art. 3(1) lists 'sostenibilità' (sustainability) among the binding general principles governing AI development and use, alongside transparency, proportionality, security and non-discrimination. No operative environmental-reporting or training-footprint duty. |
| UN Global Digital Compact | implicit | Promote sustainability across the life cycle of digital technologies ...; potential negative impacts of emerging digital technologies on ... the environment. (paraphrase) GDC para 11(e) lifecycle sustainability; Objective 5 narrative (A/RES/79/1, Annex I) |
Evidence convergence
Sources the corpus cites for this theme across multiple articles — a scientometric consensus signal computed from inline prose citations (the more articles independently cite a source, the more load-bearing it is for this theme). 25 sources are cited by ≥2 articles.
- 8×AI, Climate, and Regulation: From Data Centers to the AI Act — cited by 8 articles
- 6×An interdisciplinary account of the terminological choices by EU policymakers ahead of the final agreement on the AI Act: AI system, general purpose AI system, foundation model, and generative AI — cited by 6 articles
- 6×A Framework for Evaluating Global AI Governance Initiatives — cited by 6 articles
- 5×Missing the Mark: Adoption of Watermarking for Generative AI Systems in Practice and Implications Under the New EU AI Act — cited by 5 articles
- 5×Large language models reflect the ideology of their creators — cited by 5 articles
- 4×European ambitions captured by American clouds: digital sovereignty through Gaia-X? — cited by 4 articles
- 4×The simple macroeconomics of AI — cited by 4 articles
- 4×Artificial intelligence and synthetic biology: biosecurity risks, dual-use concerns, and governance pathways — cited by 4 articles
- 4×The EU model of AI governance: regulating artificial intelligence through law and policy — cited by 4 articles
- 3×arxiv:2504.18236 — cited by 3 articles
- 3×Audio deepfakes and the regulation of the landlords of creativity — cited by 3 articles
- 3×Generative AI at Work — cited by 3 articles
- 3×Generative AI and data protection — cited by 3 articles
- 3×Identifying Algorithmic Decision Subjects' Needs for Meaningful Contestability — cited by 3 articles
- 3×Making AI Less 'Thirsty': Uncovering and Addressing the Secret Water Footprint of AI Models — cited by 3 articles
- 3×The Current Landscape of Deepfake Legislation in the United States — cited by 3 articles
- 2×Infrastructure for AI Agents — cited by 2 articles
- 2×Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law (Council Eur.) — with Introductory Note — cited by 2 articles
- 2×Multi-Agent Risks from Advanced AI — cited by 2 articles
- 2×The establishment of an international AI agency: an applied solution to global AI governance — cited by 2 articles