A Conversational Agent for Argument-driven E-participation.

The majority of current e-participation tools are based on online web forums where citizens make proposals and provide comments and opinions, forming large conversation threads. Motivated by the huge popularity of instant messaging applications and the impressive, recent advances in natural language processing and artificial intelligence, in this paper we propose and investigate the use of conversational agents or chatbots as a new form of citizen-to-government communication. Specifically, we present and evaluate a novel chatbot that assists a user on the overwhelming task of exploring the citizen-generated content of Decide Madrid, a forum-based e-participatory budgeting platform. Among other things, the proposed chatbot is capable of automatically extracting, categorizing and summarizing the arguments underlying the citizen proposals and debates in the platform. Through a user study, we show promising results about the potential benefits of the chatbot in terms of several citizen participation, decision making and public value criteria.