Green AI? A Critique of Green Techno-Optimism

This article interrogates one of the latest manifestations of the ideology of “green techno-optimism”: the belief that new digital technologies—now artificial intelligence (AI)—can and will save the planet. Drawing on political economy and energy humanities, we probe the new discourse of “green AI” which is emerging across corporate, governmental, and media spheres. While this discourse represents AI platforms and apps as capable of addressing the problem of global climate change, it also obscures their material, labor, and ecological costs. We thus critique the green techno-optimist ideology surrounding AI today in support of a greater “green tech consciousness” and “green tech conscience.” We conclude the essay with “nine laws of AI-ecology,” inspired by the “Four Laws of Ecology” Barry Commoner (1979) introduced in The Closing Circle. These distill our critique of green AI into a useful heuristic for reading what AI does to the world, and what it asks us to ignore. These laws are: (1) everything is connected to everything else; (2) everything must go somewhere; (3) nature knows best; (4) there is no such thing as a free lunch; (5) there is no infinite energy on a planet of finite resources; (6) efficiency increases consumption; (7) planned obsolescence negates efficiency gains; (8) efficiency cannot mitigate the environmental consequences of economic growth; and (9) all “green” is greenwashing. The most important task is not to make machines more intelligent in the hope that they will do the labor required to save the planet for us, but to make humans more intelligent about the true planetary consequences of AI systems, so they may find more environmentally responsible and sustainable ways of making and using AI.