ChatGPT Struggles with Rural Environmental Justice

ChatGPT addresses rural environmental justice while minimizing its ecological footprint

In the era of advanced technology, artificial intelligence, and language models like ChatGPT, it's crucial to explore how these innovations intersect with issues of environmental justice in rural areas. While ChatGPT is a remarkable tool, its limitations in addressing rural environmental justice concerns are evident.

A recent study conducted by researchers at Virginia Tech University in the United States delved into the capabilities of ChatGPT, an artificial intelligence tool developed by OpenAI, in providing information on environmental justice issues in rural counties across the country. The researchers aimed to assess the AI model's performance by prompting it to address environmental justice concerns in each of the 3,108 counties in the United States.

ChatGPT, a large language model trained on extensive natural language data, demonstrated proficiency in identifying location-specific environmental justice challenges in large, high-density population areas. However, the study uncovered limitations when it came to addressing local environmental justice issues in rural counties. Out of the total 3,018 counties queried, ChatGPT could only provide location-specific information for approximately 17 percent, specifically 515 counties. These findings were published in the journal Telematics and Informatics.

Junghwan Kim, an assistant professor at Virginia Tech University and the corresponding author of the study, emphasized the importance of investigating the technology's limitations to ensure awareness of potential biases among future developers. The driving motivation behind the research was to shed light on the boundaries of generative AI tools and their susceptibility to biases.

The study revealed notable geographic biases in the ChatGPT model. In rural states like Idaho and New Hampshire, over 90 percent of the population resided in counties where the AI tool could not provide local-specific information. Conversely, in states with larger urban populations such as Delaware or California, less than one percent of the population lived in counties without access to specific information.

The researchers underscored the need for further investigation, emphasizing that the findings highlighted concerns about the reliability and resiliency of large language models. Ismini Lourentzou, a study co-author from the Department of Computer Science, pointed out potential issues regarding the robustness of these models.

ChatGPT’s environmental impact is a reminder that we need to be more mindful of the environmental costs of AI development. We need to develop sustainable practices that reduce the carbon footprint of AI models like ChatGPT. We also need to ensure that the benefits of AI are accessible to everyone, including rural communities