We were inspired by the idea of combining GPT-4 with a Legal Database to help legal professionals cut down on the time and resources they spend on legal research.
What it does
The lawyer using LawGPT uploads the files associated with the civil case they are working on, and LawGPT uses GPT-4 to create summaries of the key points of each document. It then uses these summaries to find inconsistencies in the information contained, suggest additional documents that could help the lawyer build a stronger case, and searches WestLaw for similar cases which it summarizes and provides to the user.
How we built it
We used GPT-4 and langchain to implement our AI, tesseract to transcribe pdfs, flask for our web server, selenium for our web scraper, and react for our frontend.
Challenges we ran into
We are both new to React, so that presented the majority of our challenges. Besides that, since WestLaw doesn't have an API (that we could use) we had to build a web scraper to interact with the platform instead which also presented some unique challenges.
Accomplishments that we're proud of
We're proud of the fact that we completed all the features we set out to build to a functional point; the AI fully summarizes the documents provided and uses WestLaw to conduct additional case law research, which was our biggest stretch goal initially.
What we learned
We both learned a lot about building React apps.
What's next for LawGPT
Hopefully we can continue to build on this idea of an AI-powered document processing assistant in the future.