The Environmental Paradox of Artificial Intelligence

Rethinking Liability and Sustainability in Algorithmic Governance

Abstract

Artificial Intelligence (AI) is often framed as a driver of environmental sustainability, enabling advances in climate science, conservation, and resource efficiency. This paper challenges that view, arguing that AI is also an emerging source of environmental harm that current legal systems fail to address. Its energy-intensive computation, expanding data infrastructures, and extractive supply chains embed ecological costs within supposedly sustainable technologies. A core regulatory gap underpins this paradox. Environmental law, designed for identifiable polluters and linear harms, struggles with AI’s diffuse, opaque, and transnational impacts. Meanwhile, AI governance focuses on ethics and data, largely ignoring environmental concerns, creating a “liability vacuum.” This gap is evident in Digital Sequence Information, where AI-driven genomics bypasses existing biodiversity frameworks, raising issues of equity and sovereignty. The paper calls for re-centering environmental principles in AI governance, extending doctrines like precaution and polluter-pays to address distributed technological harm.