Enhancing structured democratic dialogues: evaluating the impact of AI assistants on participant contributions
|
Abstract
The structured democratic dialogue (SDD) process involves distinct stages designed to enhance participants’ understanding and refinement of each other’s contributions. Initially, participants generate titles for their responses to a predetermined triggering question, followed by brief clarifications. As the dialogue progresses, further exploration of ideas necessitates additional elucidation, deepening their meaning and concreteness. This paper investigates the integration of an AI-enriched model into the process, through a Python-based program utilising ChatGPT-4, named AI-SDD-Assistant©. The study evaluates the AI-assistant’s effectiveness in enhancing participants’ original contributions by making them specific, measurable, assignable, relevant, and time-bounded (SMARTer). Participants utilised the AI-assistant to refine their ideas based on clarification questions from others in subsequent stages. Results indicate that the majority of participants approved the AI-assistant’s enhancements, resulting in more specific, measurable, assignable, and relevant responses, with time-bound estimates. While one-third of the participants remained skeptical, they still endorsed the use of AI-assisted revisions by organisers.