Artificial intelligence and the regulation of the social metabolism. On the interaction between AI and the biosphere

This article presents a framework for understanding the role of artificial intelligence (AI) in the ecological crisis. AI is conceptualized as an epistemic infrastructure that regulates the social metabolism with the biosphere, primarily through techniques of prediction. However, the relationship between AI and the biosphere is not one of unilateral control but rather one of interaction, conceptualized as iterative processes of externalization and internalization. The article identifies two mechanisms accounting for the material dimension of this interaction and two mechanisms accounting for its symbolic dimension. The first mechanism addresses the material externalization of the epistemic infrastructure, arguing that it does not suffice to account for the direct resource demands of AI, emphasizing indirect ecological effects that occur through the acceleration of production and circulation. The second mechanism highlights AI's symbolic externalization in environmental governance. The third mechanism addresses the material dimension of the internalization of biospheric elements into AI. This happens when the development of the epistemic infrastructure has to be adjusted to the materiality of the biosphere. The fourth mechanisms addresses the symbolic internalization of biospheric elements into epistemic infrastructures through biomimetic design. The framework is developed by tracing the historical evolution of digital technology into AI. The article concludes by emphasizing that the predictive capacities of AI used to coordinate the expansion of the social metabolism have also become a core element in in the production of knowledge on the ecological consequences of this expansion, thereby offering a potential sensitizing apparatus for responding to the ecological crisis.