Legal Reasoning AI
The mission of this project is to develop an AI agent capable of performing legal reasoning within the Swiss private law ecosystem. Imagine an interface where legal experts can pose questions and, in return, receive step-by-step explanations of the process leading to a legally correct answer. The first step will highlight relevant legal sources identified by a semantic search algorithm that sifts through curated law articles, while the following process will involve LLM queries that synthesize these sources into a concise response. The advantage of such a system lies in its ability to automate legal reasoning, offering legal professionals full transparency to verify the legal basis of the answers. To achieve this, we utilize the existing RAG (Retrieval-Augmented Generation) solution from the Law Bot open-source software project, which provides a robust technical foundation that can be further developed. Contributors to the Law Bot project will provide guidance on enhancing this existing solution. This challenge represents an opportunity to blend AI with legal expertise, ensuring both accuracy and transparency in resolving use cases in the area of Swiss private law. We invite you to join us in pushing the boundaries of legal technology, making legal reasoning both automated and verifiable.
Law Bot
Backend
The Backend is based on Python / FastAPI and described in more detail in the backend folder.
Frontend
The Frontend is based on Next.js and described in more detail in the frontend folder.
Data
Test data can be found inside the data folder.