Improving citizen-government interactions with generative artificial intelligence: Novel human-computer interaction strategies for policy understanding through large language models
2024

Improving Citizen-Government Interactions with Generative AI

Sample size: 100 publication 10 minutes Evidence: moderate

Author Information

Author(s): Yun Lixin, Yun Sheng, Xue Haoran

Primary Institution: School of Humanities and Social Sciences, Qingdao Agricultural University, Qingdao, Shandong, China

Hypothesis

Can integrating Retrieval-Augmented Generation (RAG) technology with Large Language Models (LLMs) enhance the communication of government policies to citizens?

Conclusion

The proposed system significantly improves the accessibility and comprehensibility of government policies, fostering better citizen engagement.

Supporting Evidence

  • The system demonstrated high accuracy, averaging 85.58% for Chinese and 90.67% for US policies.
  • Evaluation metrics included accuracy, comprehensibility, and public engagement.
  • Case studies highlighted the system's impact on transparency and citizen interaction.

Takeaway

This study shows that using advanced AI can help people understand government rules better, making it easier for them to ask questions and get answers.

Methodology

The study utilized a system combining Retrieval-Augmented Generation (RAG) technology with Large Language Models (LLMs) to process and respond to citizen queries about government policies.

Potential Biases

Potential biases may arise from the specific datasets used, which could affect the generalizability of the findings across different regions and policy areas.

Limitations

The system's performance varied between Chinese and US policy documents, indicating a need for further tuning to accommodate linguistic and structural complexities.

Participant Demographics

The dataset included policy documents from both Chinese and US regional governments, covering various policy topics.

Digital Object Identifier (DOI)

10.1371/journal.pone.0311410

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication