Improving citizen-government interactions with generative artificial intelligence: Novel human-computer interaction...

Area of Science:

• Public Administration
• Artificial Intelligence
• Information Science

Background:

• Effective government policy communication faces challenges like complexity and accessibility.
• Digital transformation and Generative AI offer new avenues for public administration.
• Citizen engagement is vital for transparency and democratic processes.

Purpose of the Study:

• To develop and evaluate a system using Retrieval-Augmented Generation (RAG) and Large Language Models (LLMs) to improve government policy communication.
• To address challenges in policy accessibility, comprehensibility, and citizen engagement.
• To enhance the dynamic interaction between public authorities and citizens.

Main Methods:

• Implementation of a system combining Retrieval-Augmented Generation (RAG) with Large Language Models (LLMs).
• Utilizing a sophisticated retrieval mechanism to generate accurate and comprehensible responses to citizen policy queries.
• Experimentation with a diverse dataset of over 200 policy documents from Chinese and US regional governments.

Main Results:

• The system achieved high accuracy rates: 85.58% for Chinese policies and 90.67% for US policies.
• Evaluation metrics included accuracy, comprehensibility, and public engagement, showing significant improvements.
• Case studies demonstrated the system's positive impact on transparency and citizen interaction.

Conclusions:

• The integrated RAG and LLM system significantly enhances policy communication accessibility and understandability.
• The approach effectively boosts public engagement, contributing to a more informed and participatory democratic process.
• This innovative method represents a substantial advancement over traditional policy dissemination methods.