Artificial intelligence’s hidden environmental and public-health risks: Should AI be considered a public health threat?

The rapid worldwide growth of artificial intelligence (AI), especially generative models, has intensified energy demands, contributing to emissions, water depletion, and e-waste. Beyond environmental pathways, AI poses public health risks through algorithmic bias, health misinformation, mental health impacts, and labor displacement. This commentary synthesizes evidence on adverse effects across the AI lifecycle through an environmental justice lens. The analysis reveals several critical issues: data center electricity demand may surpass 1,000 TWh yearly by 2026, training individual large models requires hundreds of thousands of liters of water, toxic e-waste predominantly affects developing regions, biased algorithms worsen healthcare disparities, and AI-generated misinformation diminishes public confidence. Disadvantaged populations and the Global South face disproportionate harm with minimal benefit sharing. Current regulatory frameworks are fragmented. Mitigation requires transparency, renewable energy transition, circular economy principles for hardware, health impact assessments, and robust policy. True AI sustainability must encompass ecological integrity, public health protection, and environmental justice. However, realizing these mitigation strategies faces significant feasibility challenges related to cost, technical infrastructure, and governance, which must be addressed through detailed roadmaps and economic analyses.