This article was published on November 1, 2017

AI fought the law, and won

AI fought the law, and won Image by: CaseCruncher
Inés Casserly
Story by

Inés Casserly

An artificial intelligence lawyer, CaseCruncher Alpha, just beat 100 lawyers in a challenge.

The robot and 100 of London’s top lawyers were put up against each other and given “the basic facts of hundreds of PPI (payment protection insurance) mis-selling cases and asked to predict whether the Financial Ombudsman would allow a claim.”

In total, they submitted 775 predictions, and as it turns out, CaseCruncher had an accuracy rate of 86.6 percent, while the students only got 66.3 percent correct. But according to CaseCruncher, machines are only better at humans in predicting outcomes when the question is defined “precisely,” meaning: we have a long way to go.

CaseCruncher is the project of three Cambridge Law students: Jozef Maruscak, Rebecca Agliolo and Ludwig Bull. At first it was a simple chatbot that answered legal enquiries, but was later developed to predict the outcome of court cases. None of the students have a background in computer science or technology, but were always keen on the topic.

Machines like these pose the question as to whether artificial intelligence will replace real-life attorneys. For now they can only analyze information they receive, lacking human touch and other virtues needed in a lawyer.

The BBC posed the question whether this will “be a tool for junior lawyers to use or something which replaces them.”

But for CaseCruncher, “The use case for systems like CaseCruncher Alpha is clear. Legal decision prediction systems like ours can solve legal bottlenecks within organizations permanently and reliably.”

Get the TNW newsletter

Get the most important tech news in your inbox each week.

Back to top