Blog

Predictive Coding in the Courts?

In her recent Law Technology News article, “Robot Review,” Farrah Pepper assesses the state of the “predictive coding” world. The article concisely addresses the range of issues at the forefront of the use of automated review (i.e. benefits versus advantages over manual review, including speed to access key documents, costs, accuracy, etc.), but appropriately and quickly gets to the heart of the matter — will the courts accept automated review?

While we all can point to helpful commentary at industry conferences, as well as in Sedona white papers and law review articles, it’s unlikely that we will see case law providing a safe harbor for the use of predictive coding any time soon. Unless, that is, there is actually a case brought to court in which a party has used automation and there is a related dispute.  So what do we do in the meantime?

First, although some negative inference is required, it is important to note that there is no case law prohibiting the use of predictive coding. This is a point I’ve heard members of the judiciary make on multiple occasions when asked, “When will we get a case specifically authorizing the use of predictive coding?”

Second, as Pepper rightly points out, without question the most consistently referenced standard both in the e-discovery case law and in the Federal Rules of Civil Procedure and the Federal Rules of Evidence is reasonableness. In our view, this is the standard that any review process — automated, manual or otherwise — should meet.

Defining or defending the reasonableness of any review is a practical task, as outlined generally in the now landmark Victor Stanley (250 F.R.D. 251, 259 (D. Md. 2008)) decision:

  • Were the people responsible for executing the review competent professionals
  • Were reasonable steps taken to ensure the quality of the process
  • Was the process, including the critical decisions made along the way, documented

When the idea of predictive coding or automated review is introduced, we believe there are at least two other important considerations that must be included in developing a reasonable or “defensible” process:

  • Is the process transparent such that the basis for any automated decision can be readily understood and explained
  • Have the parties conducted adequate sampling to ensure both accuracy rates as well as appropriate precision and recall

Our i-Decision® service is specifically designed to answer each and every one of these questions with an emphatic “yes.” Ours is a quality process, well-documented and executed by competent professionals, with complete transparency as to each and every document decision and scientifically sound statistical sampling to address even the toughest skeptic’s line of questioning.

Pepper concludes that, while there is no specific case law authorizing the use of predictive coding, “the buzz is just starting.”  Our conclusion is that we are well beyond buzz.  The tools to undertake automated review are here.  And in the hands of competent professionals with well-defined processes, predictive coding, in the very near future, will set the benchmark for reasonableness.

Maureen O'Neill