“Big Data” is one of the most important trends in technology, and now it’s making waves in the legal community because of its potential to predict legal outcomes and costs for specific cases. Traditionally lawyers have to answer the seemingly straightforward question, “How much is this going to cost me?” with a combination of instinct and guesswork, but software to accurately predict costs using historical data is in development. It will likely be some time before any type of functional software becomes available for every field of law because getting enough usable data is extraordinarily difficult, but the possibility is intriguing nevertheless.
Despite still being in its infancy, the field of automated trial outcome prediction does have some players that are already building up the field. Lex Machina, a start-up that has been building and organizing a database focused solely on IP law for the last ten years, may be the furthest along in producing a usable product. Right now their database consists of 128,000 IP cases and many thousands of attorney records, judges, law firms, and parties spanning the last decade. However, the last mile of analysis still requires human thought and input.
Perhaps most fascinating is how this could eventually lead to something a lot like insurance actuarial tables for litigation. For an area as complex as trial and settlement outcomes this is very impressive and could have huge implications for the entire legal field. It is conceivable that a bot could crawl through data sources like PACER (Public access to court electronic records) to create a database of Big Data to use in this type of prediction, but the many challenges involved in using those records indicate that this will likely be a distant reality. Most of the potential sources for Big Data for even very specific types of law have huge numbers of incorrectly coded documents and case outcomes, which at present requires engineers and data analysts to spend hundreds of thousands of hours cleaning up and organizing documents to make them usable.
However far off in the future this reality may be, we know that an existing computer model is already better, on average, than legal experts at predicting the outcome of supreme court cases. While the Supreme Court obviously accounts for a very small segment of cases in the United States, the implications of this type of better-than-human prediction for other fields of law are far-ranging. Combined with rapidly accelerating tools to predict potential legal costs, companies and firms will be able to make much more informed decisions on the most cost-effective way to pursue a matter.