Artificial intelligence (AI) is rapidly transforming medical imaging, offering new opportunities to improve diagnostic accuracy, accelerate workflows, and optimize resource use. Yet alongside these clinical benefits, AI also introduces significant environmental costs. Training and deploying advanced models require substantial computational resources, while the manufacture and operation of imaging hardware, data centers, and supporting infrastructure generate greenhouse gas emissions, consume electricity and water, and contribute to electronic waste. With medical imaging already responsible for an estimated 1% of the global carbon footprint, the integration of AI presents both a risk of exacerbating this burden and an opportunity to mitigate it. This chapter examines the dual role of AI as both a potential contributor to and a solution for healthcare’s environmental impact. We discuss how AI can reduce low-value imaging, shorten scan protocols, improve dose efficiency, and lower reliance on environmentally harmful contrast agents. AI-driven triage, scheduling, and telemedicine approaches can reduce scanner idle time, improve patient flow, and minimize travel-related emissions. At the same time, strategies are needed to reduce the footprint of AI itself. Algorithmic efficiency, energy-conscious hardware selection, sustainable data management, renewable-powered data centers, and life cycle-conscious procurement and disposal all represent critical levers for building “green AI.” We further highlight regulatory and ethical dimensions, arguing that sustainability must be recognized as a core principle in AI governance, alongside fairness, transparency, and safety. Taken together, these perspectives emphasize that sustainable AI in medical imaging requires more than technical fixes. It demands deliberate design choices, coordinated operational strategies, and multi-stakeholder governance to align clinical innovation with planetary health. When developed with environmental stewardship in mind, AI can become a powerful enabler of both high-quality care and low-carbon healthcare delivery.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Toward a Sustainable Future of AI in Healthcare Imaging

  • Saif Afat,
  • Judith Herrmann,
  • Thomas Küstner

摘要

Artificial intelligence (AI) is rapidly transforming medical imaging, offering new opportunities to improve diagnostic accuracy, accelerate workflows, and optimize resource use. Yet alongside these clinical benefits, AI also introduces significant environmental costs. Training and deploying advanced models require substantial computational resources, while the manufacture and operation of imaging hardware, data centers, and supporting infrastructure generate greenhouse gas emissions, consume electricity and water, and contribute to electronic waste. With medical imaging already responsible for an estimated 1% of the global carbon footprint, the integration of AI presents both a risk of exacerbating this burden and an opportunity to mitigate it. This chapter examines the dual role of AI as both a potential contributor to and a solution for healthcare’s environmental impact. We discuss how AI can reduce low-value imaging, shorten scan protocols, improve dose efficiency, and lower reliance on environmentally harmful contrast agents. AI-driven triage, scheduling, and telemedicine approaches can reduce scanner idle time, improve patient flow, and minimize travel-related emissions. At the same time, strategies are needed to reduce the footprint of AI itself. Algorithmic efficiency, energy-conscious hardware selection, sustainable data management, renewable-powered data centers, and life cycle-conscious procurement and disposal all represent critical levers for building “green AI.” We further highlight regulatory and ethical dimensions, arguing that sustainability must be recognized as a core principle in AI governance, alongside fairness, transparency, and safety. Taken together, these perspectives emphasize that sustainable AI in medical imaging requires more than technical fixes. It demands deliberate design choices, coordinated operational strategies, and multi-stakeholder governance to align clinical innovation with planetary health. When developed with environmental stewardship in mind, AI can become a powerful enabler of both high-quality care and low-carbon healthcare delivery.