Artificial intelligence (AI) in radiology presents a dual challenge: it contributes significantly to greenhouse gas (GHG) emissions due to energy-intensive model development and data storage, but also offers opportunities to enhance environmental sustainability through optimized imaging protocols, reduced scan times, and efficient resource use. The article discusses the environmental impact of AI in radiology, emphasizing the need for strategies to minimize emissions while leveraging AI to improve sustainability. Key challenges include the high energy consumption of AI model training and inference, the environmental impact of data storage, and the need for sustainable energy sources. The article highlights the importance of energy-efficient configurations, cloud-based solutions with renewable energy, and collaborative efforts to reduce redundancy and improve model validity. It also explores how AI can optimize image acquisition, reduce contrast agent use, and improve patient scheduling to lower emissions. The article concludes with a call for research, education, and policy changes to ensure AI in radiology contributes to environmental sustainability while maintaining clinical effectiveness. The top 10 actions to improve sustainability include using energy-efficient AI configurations, developing GHG emission calculators, encouraging collaboration, optimizing data storage, reducing scan times, and using AI for clinical decision support. The article underscores the need for a balanced approach to AI in radiology that addresses both environmental and clinical benefits.Artificial intelligence (AI) in radiology presents a dual challenge: it contributes significantly to greenhouse gas (GHG) emissions due to energy-intensive model development and data storage, but also offers opportunities to enhance environmental sustainability through optimized imaging protocols, reduced scan times, and efficient resource use. The article discusses the environmental impact of AI in radiology, emphasizing the need for strategies to minimize emissions while leveraging AI to improve sustainability. Key challenges include the high energy consumption of AI model training and inference, the environmental impact of data storage, and the need for sustainable energy sources. The article highlights the importance of energy-efficient configurations, cloud-based solutions with renewable energy, and collaborative efforts to reduce redundancy and improve model validity. It also explores how AI can optimize image acquisition, reduce contrast agent use, and improve patient scheduling to lower emissions. The article concludes with a call for research, education, and policy changes to ensure AI in radiology contributes to environmental sustainability while maintaining clinical effectiveness. The top 10 actions to improve sustainability include using energy-efficient AI configurations, developing GHG emission calculators, encouraging collaboration, optimizing data storage, reducing scan times, and using AI for clinical decision support. The article underscores the need for a balanced approach to AI in radiology that addresses both environmental and clinical benefits.