An NLG pipeline for a legal expert system: a work in progress

TL;DR

This paper describes the development of a natural language generation component for a legal expert system, which transforms structured legal data into human-readable text through a two-step pipeline.

Contribution

It introduces a novel NLG pipeline integrated into a domain-specific language for legal drafting, demonstrating its application in a legal expert system.

Findings

The NLG component effectively generates natural language from legal data.

The pipeline supports legal interviews and reasoning outputs.

Initial implementation shows promising results for legal document drafting.

Abstract

We present the NLG component for L4, a prototype domain-specific language (DSL) for drafting laws and contracts. As a concrete use case, we describe a pipeline for a legal expert system created from L4 code. The NLG component is used in two steps. The first step is to create an interview, whose answers are processed into a query for an automated reasoner. The second step is to render the answers of the reasoner in natural language.