City law firm Berwin Leighton Paisner (BLP) has won what it believes to be the first contested application to use predictive coding as part of a substantial document review exercise.
Predictive technology – which learns from human document reviewers and prioritises and categorises documents automatically using algorithms – is growing in importance, with Master Matthews allowing its use in England and Wales for the first time earlier this year, in the case of Pyrrho.
Already in use in America, the Irish courts gave the technology their approval last year.
According to BLP, the case is an unfair prejudice petition in which the petitioner is seeking a buy-out of his minority shareholding. The respondents strongly contest the allegations, and the valuation suggested by the petitioner. The parties agreed most case management directions in advance of the first case management conference, with the respondent’s approach to disclosure the most substantial point of dispute.
BLP said the respondent’s preference for a ‘traditional’ approach to document review would be excessively costly, and that superior results could be achieved at a more proportionate cost using predictive coding technology.
It was estimated that a traditional linear review would cost more than two and a half times the cost of predictive coding, partly because BLP is unusual in having in-house data processing, hosting and document review capabilities, and – it said – “virtually unique” in having in-house predictive coding technology.
BLP argued that almost all of the factors set out by Master Matthews in favour of using predictive coding applied in this case, with the lack of agreement between the parties the main one that did not. The court agreed and ordered the respondent’s solicitors to use predictive coding.
The firm said the ruling was a clear demonstration that predictive coding technology was not only suited to handling claims with exceptionally large datasets – there were three million documents in Pyrrho – but could also be “critical” to achieving proportionality in smaller claims involving more typical datasets (there are approximately 500,000 documents in this case).
Lead partner Oliver Glynn-Jones, head of commercial dispute resolution at BLP, said: “The technology will not only reduce the cost of e-disclosure, but also operates at a higher level of accuracy than a traditional human review. It also opens up new opportunities such as early case assessment, since it enables lawyers to quickly identify the most highly relevant documents at a much earlier stage than through a traditional review.”