Artificial intelligence lawyer called “CaseCruncher” proved that it is much superior to human lawyers in predicting winnability of legal cases.
In a rare match-up held last month, over 100 lawyers from London’s top-notched law firms competed against CaseCruncher on who best can predict the winnability of legal cases.
Both AI lawyer and human lawyers were given basic facts about legal cases pertaining to payment protection insurance (PPI) mis-selling cases and they were asked to guess whether the Financial Ombudsman would allow the claims in the mis-selling cases. The competition was overseen by two judges, Cambridge law lecturer Felix Steffek and Ian Dodd from Premonition – a company that operates the world’s biggest database of legal cases.
Both CaseCruncher and human lawyers submitted a total of 775 predictions. The AI won, with CaseCruncher giving an accuracy rate of 86.6%, compared with 66.3% for the human lawyers.
CaseCruncher is the brainchild of three Cambridge law students, who built their AI starting from a chatbot that answered legal questions.
“A CaseCruncher is a system that makes legal decision predictions,” the three humans behind CaseCruncher described their AI lawyer. “CaseCrunchers are for organisations looking to remove legal bottlenecks permanently. The final result is a system that reduces time and money spent on a legal problem while matching or outperforming human accuracy.”
Dodd from Premonition told the BBC that AI may replace some of the work done by paralegals and junior lawyers. He said, however, that no AI can talk to a client or argue in front of a High Court judge.