Did You Remember To Plug In Your Lawyer Today?
Ryan Etter/Getty Images
Is it time to start preparing for the robot invasion?
Dentons, via its subsidiary NextLaw Labs, announced a joint venture this month with a Toronto-based startup that is developing a new artificially intelligent legal assistant, called ROSS, that can research case law and answer legal questions posed in plain English.
ROSS Intelligence Inc.’s 26-year-old CEO, University of Saskatchewan law grad Andrew Arruga, says he and a few tech-savvy friends saw an opportunity to apply the growing power of AI to legal research, an area that he says costs clients about $8.4 billion annually in the United States.
The team began by feeding thousands of pages of court and legislative documents related to Canadian labor and employment case law to a version of Watson, IBM’s cognitive computer. Then they trained it to mine the database to infer answers to legal questions. In mid-May, ROSS entered the U.S. with a new AI creation aimed at answering insolvency case law questions.
The machine is learning much the way a junior associate does, Arruga says, with its opinions reviewed and critiqued by a group of law partners who signed on to beta-test the technology.
Before too long, it will “kind of be like having a senior partner in your pocket 24 hours a day,” Arruga predicts. Expansion into other practice areas is in the pipeline.
Arruga’s enthusiasm got us wondering: Should young associates, or even partners, be worried about their talents becoming obsolete as artificial intelligence continues to improve? For answers, we spoke with Stowe Boyd, a self-described futurist and research director for Gigaom Research who focuses on technology and the evolution of work. He offered his own take on recent advances in legal AI and how they’re likely impact the industry.
Am Law Daily: Ross Intelligence has called its new product an “artificially intelligent attorney.” Are they offering something really revolutionary here?
Stowe Boyd: They’re one of a number of parallel efforts to innovate in this area. There’s a lot of work already being done by algorithmic software that’s been around for quite a while. The work for paralegals has in some cases been significantly reduced because of the availability of these tools. They’re getting better. Instead of a lawyer having to put in key word searches, the newest-generation algorithmic software can come up with what is lying buried within documents.
I saw a demo of a product by the tech company Brainspace where they parsed millions of unsorted, unstructured emails on a server from the Enron case that back in 2000 took reviewers months to parse through. Using the program, 20 minutes later the software told them what the emails were about. It was: “here’s a set of emails where people are explicitly talking about illegal things.” Instead of many months and many document reviewers, one person could do an analysis in a half hour and get to the heart of the issue.
The breakthrough is you don’t have to come up with classifications and taxonomy in advance—the algorithm derives the classifications by looking at the documents. So it minimizes the involvement of human beings in organizing the information.
The reality is that law, like almost any professional services industry that currently requires a human being to read and analyze massive amounts of data, is too expensive. Innovations like this one make a lot of the work of e-discovery experts obsolete.
ALD: How is what ROSS is doing different?
Boyd: Brainspace is based on the analysis of text. The Watson-based system and all of its new manifestations, including ROSS and medical diagnostic systems, are building knowledge databases that are able to understand natural language. They can weigh a lot of factors and come up with a prognosis. Human beings are really bad at that. Human beings boggle after the 13th or 15th or 18th factor to consider. We have to use mental shortcuts. A machine like Watson doesn’t. It can be taught to understand the complex interactions of many, many factors. And it can learn from its mistakes.
ALD: So it looks like lawyers will need to do a lot less document review and research on legal questions in the future. What will lawyers be doing then? How senior up the professional food chain can AI go?
Boyd: Well, there is a venture capital firm that recently put an algorithm program as a member of its board. The program will vote on whether to invest in specific companies or not. But in the near term, it’s unlikely we’ll see artificial intelligence taking over the position of a senior partner of a law firm. Five to 10 years from now, however, those who advance to partnership may be those who figure out how to best apply AI to the work they’re involved in.
There’s an analogy to be made with freestyle chess tournaments, where humans compete assisted by machine intelligence. The winners of those tournaments are not the smartest human players or the fastest computers, but a combination of both. Think about a freestyle lawyer of the future: He or she may not be the best lawyer in today’s terms, but he may be a person who understands both the law and how to use computers to work through complex problems. In that way they’ll be able to take advantage of the differences in how human beings and artificial intelligence attack problems.
Source: Did You Remember To Plug In Your Lawyer Today?
Via: Google Alerts for AI