Scientists and decision-makers recognise artificial intelligence (AI) as a potential tool to solve sustainability issues. However, AI solutions can be leveraged for different ends and through different means. This paper presupposes that this process is contingent upon overarching environmental security discourses. It reviews how the use of AI as presented in the literature fits into national, international, human and ecological security perspectives [1]. A national climate security discourse could emphasise military uses of AI and its role in propaganda and misinformation. The international security discourse suggests that international organisations can take advantage of AI to conduct their mission. However, transnational companies can also benefit from AI, with potential negative outcomes on consumption and resource extraction. Public–private collaboration for military AI grants transnational companies and states tools to hinder environmental movements. A human security discourse emphasises the role of AI in reaching the sustainable development goals but should consider potential power imbalances to prevent inequalities. Finally, an ecological understanding of climate security emphasises the role of algorithms in shaping our vision of the environment, and how it potentially estranges us from other cosmologies and the environmental impact of AI. This reflection opens avenues to explore the interplay between AI, geopolitics and environmental protection.