Sustainable AI in medicine: navigating innovation, challenges, and environmental impact

Abstract

Background: Artificial Intelligence (AI) is revolutionizing the healthcare sector, offering unprecedented opportunities to enhance clinical efficiency, decision-making, and patient outcomes. However, alongside these benefits, AI brings considerable environmental costs due to its high computational demands. The study by Pariso et al. on AI-driven energy management in Italian hospitals illustrates the dual nature of this transformation, highlighting both the potential for increased efficiency and the structural challenges involved. In this context, the paradigm of Green AI has emerged, advocating for environmentally sustainable approaches to AI development and implementation in healthcare. Main text: While AI offers tools to optimize hospital operations, such as predictive maintenance, resource allocation, and patient flow, its widespread adoption demands vast computational resources. These requirements result in significant energy consumption, CO₂ emissions, freshwater use, and electronic waste. Data centers, essential to AI functionality, contribute notably to global electricity use and water stress, especially in areas already facing environmental constraints. To address these concerns, healthcare institutions should adopt strategies such as energy monitoring tools, lifecycle assessments, and low-carbon infrastructures. Implementing circular approaches, including waste heat reuse and equipment recycling, can further mitigate environmental impact. Beyond being a source of resource consumption, AI can also support sustainability in healthcare through intelligent systems that optimize water use, manage medical waste, and reduce material inefficiencies. Moreover, AI-enabled telemedicine, remote monitoring, and personalized patient support can significantly lower the need for physical infrastructure use, aligning healthcare delivery with environmental goals. These innovations not only support sustainability but also foster a culture of responsibility and efficiency among healthcare professionals. Conclusions: The integration of AI in medicine must be accompanied by critical reflection on its environmental footprint. Through standardized monitoring, efficient design practices, and circular resource management, healthcare systems can harness the power of AI while minimizing ecological harm. Future research should explore the environmental trade-offs of AI-enabled clinical workflows, assess energy and material use, and promote Fair AI to ensure equity and global health inclusion. By aligning innovation, fairness, and environmental responsibility, AI can fulfill its promise of advancing medical science without compromising planetary health.