AI judge found to be very accurate at predicting court case outcomes

24 Oct 2016

Digital hammer. Image: enzozo/Shutterstock

Should judges begin to feel fearful for their job security? According to a team of researchers, a new AI tool has been found to be highly accurate at predicting the outcome of European court cases.

An AI judge that computes outcomes instead of a human jury has been the dystopian creation of science fiction writers for a number of years, but new developments suggest that they could be used in some capacity in future courts.

According to University College London, a joint team of researchers from the US and the UK has developed a new AI method that is the first to predict the outcomes of a major international court, by automatically analysing case text using a machine-learning algorithm.

Judges are more realists than formalists

The international court in this case is the European Court of Human Rights (ECtHR), and by churning the data obtained during these court cases from the cases brought by defendants and prosecutors, the AI was able to predict their outcomes 79pc of the time.

The breakthrough in achieving such a high prediction rate was down to the researchers finding that judgments by the ECtHR are highly correlated to non-legal facts rather than directly legal arguments.

This suggests that judges of the ECtHR are ‘realists’ rather than ‘formalists’, based on the 584 cases that were analysed and the language used by the judges to come to their decisions.

With an equal number of violation and non-violation cases to prevent bias, the AI was found to be most confused when it came across two cases that were similar in nature, but had two different outcomes.

This, the researchers suggest, show that it is reliant on the realist language used by judges, rather than being able to pick through the finer points of the law.

‘We don’t see AI replacing judges or lawyers’

Until now, previous studies have used AI to determine whether a judge ruled in favour of or against a defendant based on the crime the person was alleged to have commit.

However, this is the first time judgments have been predicted using analysis of text prepared by the court.

“We don’t see AI replacing judges or lawyers, but we think they’d find it useful for rapidly identifying patterns in cases that lead to certain outcomes,” said the study’s lead researcher, Dr Nikolaos Aletras.

“It could also be a valuable tool for highlighting which cases are most likely to be violations of the European Convention on Human Rights.”

Before that can happen, the researchers said, the AI will need to be put through its paces again with greater amounts of case data.

Colm Gorey was a senior journalist with Silicon Republic

editorial@siliconrepublic.com