All AI risk categories
3 canonical risk pages
Environment
Energy and material externalities across the AI model lifecycle.
EwSeverity 5/10
Electronic Waste
Accelerated generation of electronic waste due to rapid obsolescence of specialized AI hardware, with environmental impact from toxic materials.
EcSeverity 5/10
Energy Cost
Significant environmental impact derived from massive energy consumption during training and inference of large-scale models, with its corresponding carbon footprint.
WtSeverity 5/10
Water Consumption
Massive consumption of water resources for data center cooling hosting large-scale AI training and inference infrastructure.