Artificial intelligence is increasingly being promoted as a tool to improve global health outcomes and tackle systemic inequities in care delivery. However, the same technologies are contributing significantly to climate change due to their massive consumption of water, energy and natural resources. Despite these impacts, the environmental costs of AI have been largely overlooked in both global health initiatives and AI ethics frameworks. This gap creates a critical paradox: technologies intended to address health challenges may, in fact, be exacerbating them. A re-evaluation of AI development is essential to align it with environmental sustainability and the foundational ethical principles of global health.
Unpacking AI’s Environmental Footprint
AI technologies demand extensive infrastructure, resulting in widespread environmental strain. Training large language models and deploying AI systems involve high electricity consumption, water usage for cooling data centres and the extraction of rare earth materials. For example, Microsoft’s water consumption rose by 34% in 2022, mainly due to its AI data operations, and global AI water demand could soon exceed that of entire countries. Similarly, the energy used by AI could rival that of nations like Sweden or Argentina. This carbon-intensive activity contributes directly to worsening climate-related health issues such as heat stress, vector-borne diseases and respiratory illnesses.
Must Read: Leveraging AI and Data for a Sustainable Future in Healthcare
Communities in the global majority—regions already burdened by weaker health systems and limited infrastructure—bear the brunt of these impacts. Data centres in areas like Querétaro, Mexico, intensify local water scarcity, while power plants tied to AI exacerbate pollution and carbon emissions. These developments mirror historical patterns of colonial extractivism, where benefits are centralised in wealthy nations while environmental and health burdens fall on already vulnerable populations. AI’s expansion is thus deepening global inequalities under the guise of technological progress.
Reframing AI Ethics for Environmental Responsibility
Existing AI ethics frameworks primarily emphasise concerns such as transparency, fairness and data privacy, often marginalising environmental sustainability. However, the environmental effects of AI should be recognised as central ethical issues in global health. A reorientation of ethical principles is necessary to address the resource consumption and climate burdens of AI.
Transparency must extend beyond algorithmic logic to encompass disclosure of resource usage, emissions and sourcing of materials. Responsibility involves not only technical accuracy but also climate accountability. Beneficence and non-maleficence, core tenets of medical ethics, demand that AI applications deliver health benefits that clearly outweigh their environmental costs. Equity and justice require a broader consideration of how the AI supply chain imposes environmental and social burdens disproportionately on the global majority. A failure to address these dimensions risks undermining the very ethical commitments AI development claims to uphold.
Advancing Climate Accountability in AI Development
Efforts to mitigate AI’s environmental impact must become embedded within research practices, funding structures and regulatory systems. Reporting standards should require the disclosure of carbon footprints associated with AI projects, including energy consumption and material sourcing. Funding bodies need to mandate climate impact assessments in grant proposals and demand mitigation plans. Algorithmic audits can be combined with environmental impact assessments to measure and manage AI’s broader consequences.
Regulatory approaches must also evolve. While initiatives such as the EU’s Artificial Intelligence Act acknowledge environmental concerns, they lack enforceable provisions. Stronger regulation is required to ensure responsible sourcing, investment in renewable energy and carbon offsetting. Furthermore, international cooperation is essential. Agreements like the Climate Neutral Data Centre Pact show promise but require broader global participation to be effective.
Incentivising environmentally conscious AI development is another critical step. Researchers and developers should be rewarded for using energy-efficient infrastructure and minimising resource use. This shift will require a cultural change, encouraging intentional decision-making that prioritises climate sustainability. Only through these combined strategies can the global health and AI sectors address their shared responsibility for environmental stewardship.
The intersection of AI, climate change and health presents one of the most pressing ethical challenges of our time. While AI holds great promise for advancing health equity, it also poses significant environmental risks that undermine this potential. Addressing the environmental impact of AI must become a core concern for global health and bioethics. This requires systemic changes across ethics frameworks, research norms and regulatory landscapes. Without such a shift, the ambition of achieving sustainable, equitable health outcomes through AI will remain out of reach. Only by fully accounting for AI’s environmental costs can its benefits be realised without compromising the health of our planet and its people.
Source: The Lancet Global Health
Image Credit: iStock
