Scholarlink

This study presents the outcomes of the conflict resolution with equitative algorithm (CREA2) project, which aims to harness the power of data engineering and AI to assist users in navigating legal jargon during the online dispute resolution (ODR) process. Starting from an initial investigation of the user-centric CREA2 platform and its scalable features, this document aims to present the human-centred information retrieval process of the built-in conversational interface. An agentic retrieval augmented generation (RAG) equipped with an advanced chain of thought prompting technique (CoT) represents the core engine of the legal AI assistant. The chatbot aims to provide coherent information consistent with trusted legal knowledge bases and responsive to user queries, while transparently displaying the data sources and motivations behind its recommendations, following a reasoning and acting (ReAct) strategy. Overall, this study outlines the user-centered architecture and design of the CREA2 platform, which aims to democratize access to justice while ensuring reliable output, serving as a blueprint for AI-driven decision support systems (DSS) in the legal sector.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A User-Centred Architecture for Legal Conversational Interfaces

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Flora Amato,
  • Mattia Fonisto,
  • Marco Giacalone,
  • Alberto Moccardi

摘要

This study presents the outcomes of the conflict resolution with equitative algorithm (CREA2) project, which aims to harness the power of data engineering and AI to assist users in navigating legal jargon during the online dispute resolution (ODR) process. Starting from an initial investigation of the user-centric CREA2 platform and its scalable features, this document aims to present the human-centred information retrieval process of the built-in conversational interface. An agentic retrieval augmented generation (RAG) equipped with an advanced chain of thought prompting technique (CoT) represents the core engine of the legal AI assistant. The chatbot aims to provide coherent information consistent with trusted legal knowledge bases and responsive to user queries, while transparently displaying the data sources and motivations behind its recommendations, following a reasoning and acting (ReAct) strategy. Overall, this study outlines the user-centered architecture and design of the CREA2 platform, which aims to democratize access to justice while ensuring reliable output, serving as a blueprint for AI-driven decision support systems (DSS) in the legal sector.