Toward an Ethics of Vulnerable Flesh in the Age of AI and the Anthropocene

Artificial intelligence is increasingly embedded in global climate governance, shaping early-warning systems, risk analytics, migration forecasts, and disaster prediction. These technologies are often framed as instruments of climate justice, promising more efficient protection of vulnerable populations through anticipatory governance. Yet AI also redistributes power by determining whose vulnerability becomes visible and whose suffering becomes actionable. Engaging debates on AI “alignment,” this paper argues that the deeper issue is not simply whose values guide technological systems but what conception of life and vulnerability they presuppose. Within climate governance, climate harms are translated into predictive models and risk indicators that render vulnerability measurable and manageable. This translation rests on an ontological assumption that life is primarily something to be secured as risk. When ontology is misconstrued in this way, ethics becomes procedural and policy becomes managerial. Drawing on phenomenology, the paper advances an ethics of vulnerable flesh that reframes climate governance around relational exposure, interdependence, and the conditions of livability.