Exploring how graduate engineering researchers approach AI sustainability: perceptions, design considerations, and impact assessment

Abstract As artificial intelligence (AI) continues to transform society, growing attention is being paid to how its own development and deployment impact the environment. This study investigated how graduate engineering researchers perceive and integrate environmental sustainability into their work. Using a mixed-methods survey of 64 graduate students at a large US research university, we explored three areas for AI sustainability: students’ perceptions of AI sustainability, the factors they consider in AI design, and how they assess the environmental impacts of their research. Our findings revealed that while students largely viewed environmental sustainability as important, the majority did not evaluate the environmental effects of their work. This gap potentially stemmed from a mix of irrelevance, lack of knowledge, and limited institutional incentives. Most students prioritized performance and research goals over environmental considerations. However, many expressed interests in learning how to integrate sustainability more effectively into their practice.