Byrom: Tensions between open justice and privacy
Access to judicial data should be made easier to increase public trust, while fears it will be used to create accurate predictions of what judges will do are overblown, a seminar heard last week.
The online Legal Education Foundation (LEF) event was accompanied by the launch of a report by academic researchers prompted by work done by LEF research director Dr Natalie Byrom, who has led calls for more justice system data [1].
It studied the handling of justice system data in three countries: Canada, Australia and Republic of Ireland.
The hope was that practices in those countries might inform policy proposals for England and Wales. But in the event, the researchers found that the situation was much the same, in that public access to data was mixed, confused and riddled with grey areas.
The LEF report, authored by Dr Judith Townend and Dr Cassandra Wiener of Sussex University, examined the use of justice system data in all its forms, from court judgments and listings information to the outcomes of cases and sentencing remarks.
While they found that Australia had made strides on the open access to data – such as pioneering what in England became the free court judgements database BAILII – and Canada had some interesting initiatives, several respondents were actually looking towards the UK for inspiration.
Dr Townend told the seminar: “[Improved] datasets would be a great enabler and facilitate better access to justice.
“It’s an important accountability check to help [such things as] public interest reporting [and] scrutiny of the system by NGOs.”
This would “help us understand how different features of the justice system impact access to justice and it would lead to greater public confidence and trust in the system.”
Another panel member, Professor Nicolas Vermeys, deputy director of the Cyber Justice Laboratory at Montréal University, said judges in Canada were concerned that open access to judicial data could become a “performance-based metric” and also encourage lawyers to engage in ‘forum shopping’ to increase chances of victory in court.
He explained: “If I can see that one judge in 95% of cases, for example divorce cases, will decide in favour of the mother and I’m representing the father… I might want to call in sick that day and ask for my case to be postponed for another day where [it can be heard by] another judge.
“Lawyers are already doing this without analytic tools [based on their experience] but this would make it easier to choose your judge based on the context of your case and could have repercussions in terms of equality of access to the courts, equality of arms et cetera.”
However, Prof Vermeys said that, very often, analysis of what individual judges had decided in the past and predictions of how they might decide in future was “extremely unreliable”.
Such litigation analytics assumed judges’ decisions were based on their own biases, and although some cases technically resulted in victory, damages may be far lower than sought, making the victory Pyrrhic.
He added that sample size was also an issue because a huge dataset was needed for artificial intelligence (AI) to be reliable: “It’s very hard to find [a large number of] examples of any one legal question, let alone linked to a particular member of the judiciary.”
Prof Vermeys concluded: “Yes, [analytics] can tell you that this judge seems to be more lenient in the afternoon than in the morning, but it’s more indicative of tendencies, [and] not necessarily something you can rely on for a decision.”
Commenting on the LEF report, Dr Byrom said: “Sharing data generated by the operation of the courts and tribunals can play a critical role in delivering open justice, improving public understanding, supporting evidence-based policy making and catalysing new legal service models.
“The digitisation of justice systems in England and Wales and worldwide… forces us to confront the tensions between open justice and privacy, which have historically been managed by the effort and expense involved in accessing information from paper-based systems.”