An artificial intelligence system has already successfully predicted the outcome of hundreds of cases at the European Court of Human Rights, according to rsearchers from the University College London and the universities of Sheffield and Pennsylvania who developed it.

According to reports, the AI “judge” examined data sets for 584 cases, with all cases either relating to torture, degrading treatment and privacy. The algorithm analyzed the English language information for each case and then made a decision – a decision that proved to be 79 percent accurate.

The vast majority of applications lodged with ECHR are deemed inadmissible, due to the fact the applications don’t meet the court’s required criteria. This means that each year the court receives thousands of applications it must read through to determine admissibility. The AI judge has been designed in part to help expedite this labor-intensive process.

“We don’t see AI jobs automation  replacing judges or lawyers, but we think they’d find it useful for rapidly identifying patterns in cases that lead to certain outcomes,” said Nikolaos Aletras, research leader at UCL’s department of computer science. For the study the AI was given both violation and non-violation cases to prove its efficacy.

“The court has a huge queue of cases that have not been processed and it’s quite easy to say if some of them have a high probability of violation, and others have a low probability of violation,” said Vasileios Lampos, also a UCL scientist and co-author of the study.

To do this, the scientists fed a database of court decisions into a natural language processing neural network. The AI then learned the outcomes after looking at certain factors including the applicants involved, where they came from, the circumstances of the case and the laws that apply to the case. Following this the AI was then given more cases and asked to determine what decision the judge would make. It was right, most of the time.

Lampos went on to say that even though the technology was promising as a filter, “laws are not structured well enough for a machine to make a decision.” Courts interpret laws differently on any given day, he added.

Photo credit: bloomsberries via Flickr