Generalized green and red concepts for artificial intelligence, big data, and information and communication technologies: Multidimensional framework and countermeasures
Organizations: School of Artificial Intelligence, Guilin University of Electronic Technology, Guilin, China · Department of Electrical Engineering, University of Chile, Santiago, Chile
Abstract
The rapid expansions of artificial intelligence (AI), big data (BD), and information and communication technologies (ICT) are driving substantial growth in electricity, water, and material demands, with data centers projected to consume 6.7--12% of U.S. electricity by 2028 (up from 4.4% in 2023) and AI-related water demand estimated at 4--6 billion m annually by 2027. Existing green AI'' and red AI'' discussions have largely emphasized computational efficiency and carbon emissions, providing limited guidance for lifecycle trade-offs, service-versus-use impacts, and broader sustainability externalities beyond energy. To bridge this gap, this paper proposes generalized green AI (GGAI) and generalized red AI (GRAI) and extends the taxonomy to red BD (RBD) and red ICT (RICT) as well as red electronic and electrical engineering (REEE) & red digital technologies (RDT) under a unified Environmental-Economic-Social-Technical/Operational (E/Ec/S/T) framework. Based on representative footprint indicators in the large-model era (electricity, carbon, water, and e-waste), we analyze how risks propagate across AI-BD-ICT stacks and across service-use boundaries. We further organize actionable countermeasures into a four-tier taxonomy aligned with the Environmental-Economic-Social-Technical/Operational framework: metric-level reform for accountability, revision of information and communication technologies design principles and demand shaping, efficiency-oriented LLM development practices (data, model, training, and scaling), and energy-infrastructure transition. The proposed framework operationalizes the Environmental-Economic-Social-Technical/Operational-aligned assessment and intervention prioritization across the AI-BD-ICT stacks, enabling sustainability-by-design without sacrificing utility.
Figures & tables
| Category | Indicator | Representative magnitude (example) |
|---|---|---|
| Energy & carbon | Data center electricity share (USA) | 4.4% (2023) projected to 6.7–12% by 2028 [ 14 ] |
| Energy & carbon | Training-related emissions (LLM example) | One GPT-3 training run reported 1.21M lb emissions [ 15 ] |
| Water | Global AI -related water demand | 4–6 billion m 3 per year projected by 2027 [ 16 ] |
| Water | On-site cooling efficiency (WUE) | Typical on-site WUE 0.2–1.8 L/kWh (site/climate dependent) [ 17 ] |
| Materials & waste | AI -driven e-waste growth (scenario) | 16 million tons by 2030 (model-based estimate) [ 18 ] |
| Concept | Scope | Primary objective | Typical evaluation cues |
|---|---|---|---|
| RAI | Model training/design | Accuracy via brute-force scaling | SOTA accuracy; larger models/data/compute |
| GAI | Model training/design | Similar utility with less compute/power | FLOPs, time, energy, CO 2 emissions |
| GRAI | System / application / use | Away from environmental, economic, social, and technical/operational sustainability goals | Net degradation in one or more dimensions (may appear efficient locally) |
| GGAI | System / application / use | Toward environmental, economic, social, and technical/operational sustainability goals | Net improvement in one or more dimensions under a stated boundary |
| Category | Near-term | Medium-term | Long-term |
|---|---|---|---|
| Metrics & accountability | Standardize footprint disclosure (electricity/carbon/water/e-waste) for training and inference; report impacts for both services and uses. | Audit-ready lifecycle accounting with region/time awareness; introduce threshold-based compliance and equity safeguards. | Institutionalize multidimensional sustainability labeling and procurement standards; enable longitudinal benchmarking. |
| System design & demand shaping | Deploy carbon-aware scheduling for non-urgent workloads; avoid peak-hour batch pipelines; reduce over-provisioned redundancy. | Capacity planning guided by demand distributions (not peak-only design); introduce service-tiering and user-side incentives. | Sustainability-by-design architectures for networks and data centers aligned with societal value and constraints. |
| Large-model development | Data curation/deduplication; selective invocation; pruning/distillation where acceptable. | Training/inference co-optimization under resource constraints; systematic multi-objective evaluation beyond accuracy. | Efficiency-first paradigms and incentives; shift away from brute-force scaling as the default development mode. |
| Energy & infrastructure transition | Increase low-carbon electricity use where available; improve cooling and water stewardship; extend hardware lifetime and reuse. | Integrate clean energy with storage and flexible loads; reduce embodied impacts via circular hardware supply chains. | Deep decarbonization of power/material/manufacturing with robust water and waste governance at scale. |
Responsible AI
Responsible AI studies how to design, evaluate, and govern AI systems to promote fairness, safety, transparency, accountability, privacy, and sustainability.
See all 132 papers