Abstract: SA-PO008
Artificial Intelligence (AI) and Green Nephrology: Assessing the Commitment of ChatGPT Models to Environmental Sustainability
Session Information
- Augmented Intelligence, Large Language Models, and Digital Health
October 26, 2024 | Location: Exhibit Hall, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Augmented Intelligence, Digital Health, and Data Science
- 300 Augmented Intelligence, Digital Health, and Data Science
Authors
- Balakrishnan, Suryanarayanan, Mayo Clinic Minnesota, Rochester, Minnesota, United States
- Thongprayoon, Charat, Mayo Clinic Health System, Mankato, Minnesota, United States
- Miao, Jing, Mayo Clinic Health System, Mankato, Minnesota, United States
- Craici, Iasmina, Mayo Clinic Minnesota, Rochester, Minnesota, United States
- Cheungpasitporn, Wisit, Mayo Clinic Minnesota, Rochester, Minnesota, United States
Background
With increasing use of AI and large language models in healthcare, including nephrology, there is growing concern about their potential impact on the environment and sustainable medical practices. Green nephrology has been increasingly emphasized in recent years due to climate change and environmental concerns. However, the commitment of AI models to green nephrology remains unclear.
Methods
Between March and May 2024, a set of 50 simulated questions, reviewed by two nephrologists, were created to assess the commitment of ChatGPT 3.5 and 4.0 to green nephrology practices. The questions were scored on a scale of 0 to 3, with 0 indicating no support for green nephrology and 3 indicating strong support. The total scores for GPT 3.5 and GPT 4 were calculated and compared. The Cohen's kappa coefficient was used to measure the agreement between the answers of GPT 3.5 and GPT 4.
Results
The total scores for GPT 3.5 and GPT 4 were 142/150 and 144/150, respectively. Both had the same score on 88% of the questions, with GPT 4 outperforming GPT 3.5 on 8% of the questions. The Cohen's kappa coefficient for the agreement between the answers of GPT 3.5 and GPT 4 was approximately 0.74. Both identified key barriers, such as high energy consumption, waste management challenges and limited awareness. They suggested strategies like implementing energy-efficient technologies, developing recycling programs and integrating sustainability into education and training to address these barriers.
Conclusion
GPT 3.5 and GPT 4 demonstrated commitment to green nephrology practices, providing valuable insights into the barriers to sustainable nephrology and potential solutions. The high agreement between them suggests that AI can be developed to support sustainable medical practices. Further research is needed on AI's environmental impact in healthcare.