In February, UK Master Matthews issued the first ruling in the EU that sanctioned the use of predictive coding as a basis for determining which electronic documents are relevant to a dispute. Since that ruling, viewed by many European legal observers as a belated approval of tools that litigators already had at their disposal, we’ve already had a case in which the High Court permitted the use of predictive coding (over the other party’s objection) as part of a substantial document review exercise.
The emergence of technology-assisted review, such as predictive coding, is certain to have a major impact on litigation throughout the EU in the months and years ahead. This was one of the topics addressed during The Cross-Border Quandary Tour, an educational program we hosted last month in London, Amsterdam and Frankfurt.
I was thrilled to be there with the world’s leading judicial authority on the topic, Judge Andrew Peck, U.S. Magistrate Judge for the Southern District of New York. Judge Peck was the first jurist anywhere to approve the use of predictive coding in his seminal DaSilva Moore v. Publicis Groupe ruling, and Master Matthews quoted extensively from Judge Peck when issuing his February decision in the UK.
“The reason predictive coding just makes sense to use is that it appears to be the best solution we have available to deal with the massive amounts of data we now have to live with in litigation,” said Hon. Andrew Peck. “In fact, research has shown that using human eyes to review litigation documents one at a time is not really the gold standard for accuracy, it’s just what we’ve done for decades because we didn’t have the technology available,” said Judge Peck. “Predictive coding leverages proven machine learning technology and the studies indicate it is more reliable for purposes of large-scale litigation document review than human eyes or keyword searching by litigation team members. Attorneys still need to review documents themselves, of course, but predictive coding has the potential of taking an initial set of 1 million emails and narrowing that down to 5,000 potentially relevant documents for attorney review.”
Judge Peck explained that best practices have emerged in the U.S. to establish a reasonable way to monitor the accuracy of predictive coding.
“We’ve learned that the best way to do quality control testing with predictive coding is to evaluate a random set of discarded documents – those that were set aside into a discard pile as being irrelevant – and then carefully inspect them one at a time,” he explained. “If we find more than a few relevant documents in the discard pile, that tells us we need to retrain the system in order to make sure it’s producing accurate results. This quality is improving in the U.S. as most of the major vendors are now producing software tools built on next-generation technology assisted review capabilities in which the systems are continually learning, retraining themselves and producing consistently accurate results in document review.”
The Cross-Border Quandary Tour focused on the challenges created as a result of conflicts between data transfer demands across global boundaries and privacy laws in the European Union. The program featured panel discussions with some of the world’s leading experts on data privacy, electronic discovery and predictive coding. To view the session featuring Judge Peck, “Predictive Coding in the UK and EU Gathering Momentum,” please click here.
Stay tuned to the AccessData Blog for additional articles based on the content from the discussions, in which attendees learned how to navigate cross-border conflict of law issues when privacy is at stake and U.S. regulators or courts demand personal data.