All work

AI/ML · 2023

Legiscan

Stanford Law / CODEX hackathon tool that compares proposed legislation to current law and scores its complexity.

Role
Developer
Timeline
Sep 2023
Team
Solo
Year
2023

Overview

A tool that compares proposed legislation against the status quo, returning a plain-language net-change summary and a complexity score.

Challenge

Bills are long and dense, and the real question is how they differ from existing law.

Solution

Used GPT-3.5 to summarize and diff legislative text and compute a readability and complexity index. Built for the Stanford Law / CODEX LLM hackathon.

Impact

  • Net-change summaries of proposed bills
  • Objective complexity scoring
  • Built at the Stanford Law / CODEX LLM hackathon

Technologies

PythonGPT-3.5NLPLangChain
Zakaria Kortam | AI Engineer & Product Engineer