May AI take health and sustainability on a honeymoon? Towards green SDTs for multidimensional health justice

Artificial Intelligence (AI) tools and machine learning methodologies used for healthcare and public health have considerably increased in the last years (Murphy et al. 2021). How should these Socially Disruptive Technologies (SDTs) be discussed from a commitment to health justice? Dealing with such a question involves not only considering the social determinants of health, but also the ecological reversals of AI used in health (Richie 2022). Despite the few existing analyses on the carbon footprint and the energy consumption of AI processes, several researchers have already warned of the potential environmental damage that may be caused if its development continues to increase (van Wynsberghe 2021, Dhar 2020). Reviewing some of these studies, we aim to rethink the global health trade-offs that AI applications for healthcare and public health might cause. Since, on the one hand, the contributions that AI applied to the health system and epidemiology may be nuanced by social, economic, political and cultural factors, and on the other hand, AI leads to environmental impacts, we wonder what is its moral balance from a concern for health justice. This involves bringing social and environmental justice into dialogue (Ausín 2021) and unravelling the contrasts of multiple health approaches. The contribution of this talk is thus threefold. First, we attempt to offer a reasonable account of health justice from a global health perspective and One Health paradigm. We consider here that the sustainability of AI-based SDTs should be addressed in the long-term and from systemic analysis, in order to assess properly the health benefits provided by them and avoiding efficiency paradoxes such as Jevons’ (York and McGee 2016). Second, when the global cost-benefit calculation in health is discussed from utilitarian lenses, we argue for including a less anthropocentric, and more relational and non-domination framework of justice to decolonise oppressive approaches to health and also to embrace ecological concerns. Finally, beyond understanding public health and global health as an unsurmountable dichotomy, efforts should be made to reconcile them when exploring the ethical impacts of AI. To this end, we conclude with a brief list of moral and justice recommendations that may serve as tipping points for advocating green and just AI applied to the field of health.