#Justicelens

Justice lens lends justice.

Comment

Inspiration= We wanted to help out the small team of pro-bono attorneys who have to handle 6000 cases yearly, and currently they only have a small team of workers that filter through the applications they receive. Not only is the application reviewing process time-consuming and inefficient, it is also vulnerable to human errors. In response to this, we wanted to develop a tool that would assist the firm in the initial screening processes to help them "weed out" the "bad apples" from the ones that really need help. They provided us a very clear goal state and we could see the impacts of our product being utilized large scale.

What it does= Our tool provides analysis on publicly available parole transcripts. It displays summaries, flags specific quotations from both the defendents and the judges based on a metric (claims of innocence from the defendents, etc.). This is intuitive, accessible to non-tech-savvy individuals. This is an integrated workspace, enabling broad-scale case management for attorneys and NPOs in the judicial field.

How we built it= We utilized Lovable as the framework for our website and incorporated Gemini as an LLM for the analysis of the court hearing transcripts. The database we use is called Supabase for backend and data management.

Challenges we ran into=Lack of publicly-available court hearing cases and documents that were behind a paywall-- if available. We had to make do with publicly available Court Parole Transcripts, which limited the extent to which we could train the language model we used.

Accomplishments that we're proud of= We linked Lawmatics -- a widely-used workspace for attourneys, one that is currenly being used by the Innocence Project -- into our website so that after analyzing a specific case through Justice Lens, it will automatically connect to the Lawmatics website. Our website provides intuitive AI for non-technical individuals to provide the most streamlined process for their work.

What we learned= maybe the real treasure was the friends we made along the way"

What's next for #Justicelens= We plan to integrate parole databases from multiple states and collaborate with legal nonprofits. We plan to talk to more experts in the legal field and gain insight into determining more specific metrics/rubrics for our "innocence score" flagging.

Built With

Share this project:

Updates