In late October, a legal skills competition between more than 100 attorneys and an artificial intelligence construct unfolded in London. More than 100 attorneys faced off against Case Cruncher Alpha, an AI construct developed and configured by Cambridge University students who initially set out to create a chatbot that can answer questions about legal topics.
According to a BBC news broadcast, the current version of Case Cruncher Alpha follows a structure that enables it to connect to a neural network and several databases. When the AI software was tasked to review 800 historic court cases involving payment protection insurance claims, it was able to accurately determine the outcome with a margin of 87 percent. The attorneys who entered the competition had less impressive results as their determination was correct by a margin of less than 62 percent.
It should be noted that the AI software had certain advantages in the competition. First of all, the lawyers were selected from various London firms, some of them quite prestigious, but they did not specialize in the type of insurance claims that they were tasked with reviewing; second, the AI was connected to several legal databases that the lawyers did not have access to when they reviewed the 800 cases.
The developers of Case Cruncher Alpha admit aware of the unfair advantage, and they admit that an experienced lawyer who specializes in payment protection insurance claims would perform better; nonetheless, the AI program took substantially less time to review a heavy caseload that required more than 100 attorneys to complete.
In the legal field, this competition seemed like the historic chess match series between IBM’s Big Blue supercomputer and Gary Kasparov in the 1990s. More recently, Google developed AlphaGo, an AI construct that beat a South Korean master in the ancient Chinese game of go, which is harder for a computer to master since playing often requires irrational strategies.
Legal experts who observed the competition stated that they think AI software may eliminate five percent of the attorney workforce in the near future.