AI Governance Fails to Address Environmental Risks

Global efforts to govern artificial intelligence must address its potential to exacerbate threats to biodiversity and ecosystems, conservationists warn, as current AI governance frameworks largely overlook environmental risks. Campaigners, including organizations such as the International Union for Conservation of Nature (IUCN), highlight that AI’s rapid development could accelerate habitat destruction, species exploitation, and resource depletion if not carefully regulated. They argue that the absence of biodiversity safeguards in AI policy discussions risks undermining global conservation goals, such as the UN’s post-2020 Biodiversity Framework, which aims to halt ecosystem loss by 2030.

Experts point to several ways AI could harm nature, including through data collection methods that disrupt ecosystems—such as satellite tracking or drone surveillance—or by enabling more efficient, but environmentally damaging, resource extraction industries. Additionally, AI-driven automation in agriculture and fisheries may intensify overexploitation of natural resources. Campaigners emphasize that AI’s energy demands, often powered by fossil fuels, further strain ecosystems through increased carbon emissions. Despite these concerns, major AI governance initiatives, including the EU’s AI Act and the UN’s AI ethics guidelines, have yet to integrate specific measures to protect biodiversity.

Conservationists are urging policymakers to embed environmental impact assessments into AI development and deployment processes, ensuring that technologies align with sustainability goals. They call for collaboration between AI researchers, governments, and environmental experts to establish binding standards that prevent AI from becoming a driver of ecological harm. With the upcoming UN Biodiversity Conference in 2024, advocates see an opportunity to pressure nations to address these gaps, stressing that safeguarding nature is essential to achieving both climate and technological progress.