The Center for Data Innovation spoke with Adam Hahn, Co-Founder and CTO of Judicata, a San Francisco-based startup working on transforming the text of legal documents into structured data and providing analytics to make it easier to conduct legal research.
Travis Korte: What’s wrong with legal research today, and how will Judicata help fix it?
Adam Hahn: The main problem can be boiled down to inefficiency. The underlying data is unstructured, so lawyers are responsible for wading through tons of data manually, which is slow and expensive. This makes certain arguments difficult or impossible to find. We add structure to this unstructured data, letting lawyers analyze the law in new ways, understand it faster, and uncover hidden insights.
TK: Structuring large amounts of legal documents sounds like a mammoth task. What are some of the things you can automate, and what are some of the things you can’t?
AH: It’s a gigantic task and will cost millions of dollars. There are few tasks we can fully automate to satisfy the accuracy a reference tool for lawyers requires, so we divide the work into automatable and human reviewed portions. For example, we automate a substantial amount of the work done to build our ontology of the law. We can find legal principles across case law in a completely unsupervised manner. Our technology then suggests relationships between this automatically derived data for our legal team to judiciously and efficiently review. Without the processes and algorithms behind these suggestions, this task would be effectively impossible.
TK: In your talk at the Wolfram Data Summit, you mentioned that you’re piloting the technology with a particular subset of case law. How did your team decide on that particular body of law?
AH: We’re beginning with California Employment Law. The reason we’re starting there is because there are many employment lawyers and a lot of employment litigation. This means there are enough cases that it’s hard for a lawyer to perform good research without our tool, and we can have a relatively large number of beta users for a small fraction of the overall case law.
TK: Once your system can handle that body of law, how easy will it be to scale up? What challenges do other specific jurisdictions pose?
AH: We’re building the technology for scaling now, and don’t anticipate any significant problems with adding other areas of law or jurisdictions. We’ve taken our time in mapping out California Employment Law because we’re doing the work of building scalable processes in parallel. You can think of it like building a car and the assembly line simultaneously.
TK: In addition to helping lawyers do legal research, I can see your technologies eventually being used to inform the public about the law. Do you anticipate making some of your data open or at least openly searchable?
AH: Yes. We think the primary data should be open and freely available. The value we add is in the analytics and more sophisticated insights we make possible. If our tools can be more widely appreciated by the public at large, we’d like to make it available to everyone.