Paper Type

ERF

Abstract

This paper proposes an LLM-assisted framework scaling citizen engagement while preserving democratic legitimacy. Unlike existing platforms focused on clustering, it provides drafting support and summarisation via human-in-the-loop architecture with modular microservices and guardrails. A proof-of-concept on real participatory budgeting data shows LLMs transform informal inputs into compliant proposals and actionable summaries (nvidia-qwen3-80B: FS=0.92, zero hallucinations). Results position LLMs as accountable civic infrastructure enhancing e-inclusion and reducing administrative burden, a key for e-democracy scalability.

Paper Number

1366

Comments

SIG E-GOV

Share

COinS
 
Aug 15th, 12:00 AM

Supporting Scalable e-Participation through LLM-Assisted Proposal Writing and Summarisation

This paper proposes an LLM-assisted framework scaling citizen engagement while preserving democratic legitimacy. Unlike existing platforms focused on clustering, it provides drafting support and summarisation via human-in-the-loop architecture with modular microservices and guardrails. A proof-of-concept on real participatory budgeting data shows LLMs transform informal inputs into compliant proposals and actionable summaries (nvidia-qwen3-80B: FS=0.92, zero hallucinations). Results position LLMs as accountable civic infrastructure enhancing e-inclusion and reducing administrative burden, a key for e-democracy scalability.