Scholarlink

This study explores the role of AI and causal inference techniques in addressing complex decision-making challenges in environmental resource management, specifically water governance. Using the DoWhy framework for causal inference, the study investigates the impact of increasing location awareness on community trust within agrarian communities managing water resources. The experiment focuses on the dynamics of farmers’ proximity to the river as a water source, comparing the effects of raising community awareness about their location. In addition, the study compares the base versions of AI models, such as ChatGPT and Claude, with those optimized through prompt engineering to address multi-stakeholder conflicts and complex environmental governance issues. The results show that while AI prompt engineering improves the model’s contextual understanding of trust dynamics, causal analysis reveals that location awareness significantly increases trust for farmers distant from the river. In contrast, it has little effect on farmers near the river. These findings demonstrate the potential of AI and causal inference to improve transparency, trust, and decision-making processes in water resource management, suggesting that targeted interventions can enhance stakeholder cooperation and inform sustainable governance practices.

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

Leveraging Large Language Models and Causal Inference to Build Trust in Environmental Decision-Making

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Suleyman Uslu,
  • Mehedi Mahmud Kaushik,
  • Praneeth Reddy Mukthapuram,
  • Sri Rama Krishna Reddy Dwarampudi,
  • Arjan Durresi

摘要

This study explores the role of AI and causal inference techniques in addressing complex decision-making challenges in environmental resource management, specifically water governance. Using the DoWhy framework for causal inference, the study investigates the impact of increasing location awareness on community trust within agrarian communities managing water resources. The experiment focuses on the dynamics of farmers’ proximity to the river as a water source, comparing the effects of raising community awareness about their location. In addition, the study compares the base versions of AI models, such as ChatGPT and Claude, with those optimized through prompt engineering to address multi-stakeholder conflicts and complex environmental governance issues. The results show that while AI prompt engineering improves the model’s contextual understanding of trust dynamics, causal analysis reveals that location awareness significantly increases trust for farmers distant from the river. In contrast, it has little effect on farmers near the river. These findings demonstrate the potential of AI and causal inference to improve transparency, trust, and decision-making processes in water resource management, suggesting that targeted interventions can enhance stakeholder cooperation and inform sustainable governance practices.