Can An AI Attorney Generate Billable Hours?

When most people think about jobs that can be lost to artificial intelligence (AI), their first though is not an attorney’s job. Case Cruncher Alpha‘s founders, Jozef Maruscak, Rebecca Agliolo and Ludwig Bull, saw the outcome of legal cases as something that deep learning could predict. The three were Cambridge law students who started out developing a chatbot to answer legal questions, even though they had no tech background.

 

To test Case Cruncher Alpha’s artificial intelligence, the U.K startup pitted its program against 100 London lawyers. Both the lawyers and Case Cruncher Alpha were given payment protection mis-selling complaints, which were already decided by a Financial Ombudsman Service. The attorneys correctly guessed the outcome of 62.3 percent of the mis-selling complaints, while CaseCruncher Alpha correctly predicted the outcome of 86.6 percent of the complaints.

 

Law firms and other organizations that currently devote resources to legal decision predictions could use Case Cruncher Alpha to do the work faster, and more accurately. Each of the AI company’s customers receive a program tailored to predict specified types of legal cases. Case Cruncher Alpha’s founders are quick to point out that the AI can make predictions when it is asked a precise question; they will be publishing a paper soon that explains exactly predictions their AI can make for organizations.

 

Case Cruncher Alpha’s AI cannot replace a flesh and blood lawyer who empathizes with clients or argues in front of a jury, however, large firms often use junior attorneys to make outcome predictions based on many billable hours of research. Would it be possible for law firms to bill for an AI’s hours or its use in the client’s case? While Case Cruncher Alpha has not released any pricing information yet, a customized AI program will not be cheap.