Canadian Lawyer

18 O C T O B E R 2 0 1 6 w w w . C A N A D I A N L a w y e r m a g . c o m T E C H S U P P O RT O P I N I O N I'll admit "legal information as data" is hardly an earth-shattering observa- tion, and it is surely over-simplistic given the dizzying array of new legal tech- nologies that do many different things in many different ways. But looking at how computers process legal information as data and turn those bits and bytes into new knowledge and insights is what I think makes legal technology the "pri- mary game changer for the legal profes- sion" (as the CBA describes it). Five innovators that I recently got to see each provide excellent examples of this. I was fortunate enough to attend the CBA Legal Conference in Ottawa this year and see The Pitch. Five legal tech start- ups were given the opportunity — in a Dragon's Den-styled event, complete with booming music and flashing lights — to "state their case". Beagle Inc, which the judges chose as the winner of The Pitch, reads contracts and highlights important aspects — such as who the parties are and their respec- tive responsibilities and liabilities, etc. It essentially turns the raw legal information (text in a contract) into data (such as the concept of a "party" with a specific value such as "Apple, Inc."). Beagle extracts these discrete provisions to display them in new ways to aid contract review and analysis for both lawyers and their clients. Loom Analytics won the "People's Choice Award" for being the most inno- vative, according to the CBA audience. Instead of contracts, this legal analytics engine turns Canadian case law into high- ly structured data. As one of their memo- rable slides had it: "Data is the New Oil." Loom uses both machine learning and legal professionals to process the decisions and apply classifications. Users are able to drill down, for example, by case type, into the decisions to see average costs and damages awards. Rangfindr was another contestant at The Pitch. It turns the salient facts used in sentencing of criminal cases into data, enabling lawyers and judges to find crimi- nal sentencing ranges "in seconds instead of hours." For example, you can search for cases that include "drinking and driv- ing offences causing bodily harm." Or you can drill down further to distinguish between bodily harm caused due to "driv- ing impaired" vs. "driving over 80," etc. Rangefindr is taking metadata associated with specific cases and turning it into an actual database that you can search, sort and filter through. Blue J Legal uses technology to simu- late how judges would decide on a novel set of facts. The simulations are based on artificially intelligent models of adjudica- tion (or machine-learning algorithms, to use the jargon) that are trained on the facts and outcomes of all the relevant decided cases in an area of law. Using specified client facts, the system produces a predicted probability of the outcome, a plain-language explanatory report and a list of closely related cases. Finally, Knomos is a new startup cre- ating a knowledge-sharing platform that have always thought of "data" as stuff to do with numbers — things you can add up, include in a formula or summarize in one of those colourful charts that spreadsheets produce. But I've been working with some legal technology recently that has challenged this notion, and it has me thinking about legal information as if it was data. When computers process legal information they can do things that we tradi- tionally do with data. They add things up, such as "how many times a particular type of clause appears in a set of contracts." They compute formulas, such as "does this set of legal arguments outweigh that set?" And they summarize legal information into a colourful chart, such as illustrating how a judge has ruled previously on specific motions or cases. Legal data: from PDFs to real answers Innovators are racing to convert legal data into much more useful formats By Kate Simpson I @k8simpson

• Cover