How a legal graduate brought AI to case preparation
Newly-qualified lawyers can spend more than a third of their time doing basic legal research. They'd prefer to do that initial work faster and more efficiently.
An MSc Computer Science student at UCL worked with an IBM Senior Inventor and IBM technology to create "Watson AI Legal Executive" (hereinafter referred to as "the app") to speed up some of that work from the outset.
A lawyer can use the app to upload previous cases to create a corpus for analysis. The app can then tell them, for their current case, which of the previous cases is most relevant, and why it's the most relevant. The app identifies similarities - for example, the nature of the alleged offence, or a pertinent Act - and provides a confidence score about the match.
The app directs the lawyer to the most relevant part, and can also provide a visual representation of the most important facts and concepts.
The app uses IBM Watson Natural Language Understanding and IBM Watson Discovery to conduct text analytics on unstructured data. This extracts entities, relationships, keywords, and semantic roles, and makes it easy to rapidly build a cognitive application in IBM Cloud.
As well as an MSc in Computer Science, the programmer also has a Bachelor of Laws degree from UCL, and was a finalist in The Privacy Law Blogging Prize. Come and see his demo, and talk to him about the potential for machine learning in law firms.
For legal reasons, we need to make it clear that all lawyers are intelligent, of course.