Skip to the content

Judges try to get their heads around AI


Michael Cross Correspondent

Get UKAuthority News


The judiciary of England and Wales has set up a new body to keep it informed of developments in the field of AI.

The AI advisory group, announced by the lord chief justice, brings together a mixture of legal, academic and policy experts to "make sure the judiciary of England and Wales is fully informed about developments in artificial intelligence".

The group is chaired by futurologist Professor Richard Susskind, who has advised successive lord chief justices on technology for more than a decade. He is also a leading figure in the drive to set up an online court for the resolution of civil disputes.

Members include one of the senior judiciary's most technophile members, Sir Geoffrey Vos, chancellor of the High Court, and the UK's leading academic expert in AI and the law, Professor Katie Atkinson of Liverpool University.

According to the announcement, the 10-strong advisory group will produce guidance on:

  • the likely impact of developments in AI on the judiciary and the court system;

  • ways of ensuring that judges are sufficiently trained on AI and its impact;

  • the most pressing legal, ethical, policy, cultural and economic effects of AI.

Lord Burnett, the lord chief justice, said: “It is vital that a modern judiciary gives systematic thought to the long term.

“AI is clearly one of the most important technologies of our day. So far, however, here and around the world, insufficient attention has been paid by judges to its impact on the work of the courts."

Accurate predictions

The move comes two years since developers on both sides of the Atlantic demonstrated that AI systems could predict the outcome of court cases with more accuracy than human legal experts.

An algorithm demonstrated by the University of Chicago successfully predicted the outcome of 70.2% of the US Supreme Court's 28,000 historic decisions. More prosaically in the UK, software developed by Cambridge start-up CaseCrunch beat human lawyers in predicting the outcome of mis-selling claims in the County Court.

No serious legal commentators are predicting that AI will make human judges redundant in the short, or even medium, term. But judges at least as likely as other senior figures on the public payroll to find their working lives changed by advances in the twin technologies of machine learning and natural language processing.

AI systems are already beginning to have an impact on commercial litigation as top City law firms deploy technology on the kind of mind numbing, evidence sifting work traditionally given to junior lawyers and trainees. This should speed up the work of the court by enabling the parties and the judge to focus on the key points at issue.

Weighing evidence

A more fundamental change could happen when AI software goes a step further, weighing up the evidence and the past performance of courts to determine what chance a case has. In the US, commercial suppliers already offer systems to help lawyers decide which courts and even which individual judges are likely to be sympathetic to their clients.

From the current situation, it is a short step to parties in a commercial dispute agreeing to let it be decided by AI, rather than going through the lengthy processes of a public court.

However, just because disputes can be settled by AI it does not mean that litigants will necessarily prefer it to having their day in court. Sir Geoffrey Vos noted last year that AI does not always take account of human frailties.

"There are many such frailties. One is unreasonableness," he told the Law Society of England and Wales. “My experience at least is sometimes that some of the most unreasonable humans can be involved in litigation.”

Image: Royal Courts of Justice by Wei-Te Wong, CC BY-SA 2.0


Register For Alerts

Keep informed - Get the latest news about the use of technology, digital & data for the public good in your inbox from UKAuthority.