Signaling the growing acceptance of predictive coding of electronically stored information - a process by which computer algorithms determine whether a document is relevant - the Delaware Chancery Court this month for the first time endorsed its use. The case, EOHB, Inc. et al. v. HOL Holdings, LLC, CA No. 7409-VCL (Del. Ch. Oct. 15, 2012), is a commercial indemnity dispute arising out of the sale of the Hooters restaurant chain. Although neither party requested that the court permit the use of predictive coding, Vice Chancellor J. Travis Laster, in a ruling from the bench, sua sponte ordered the parties not only to use predictive coding in their review of documents, but further ordered the parties to share a vendor for document hosting and review.
Although Vice Chancellor Laster has given the parties an opportunity to object to the use of predictive coding, it is not clear whether they will do so. In any event, Vice Chancellor Laster's decision is noteworthy because it appears to be the first time that the use of predictive coding, also known as computer assisted review, has been condoned by the influential Delaware Chancery Court. Computer assisted review is a process that, if used correctly, can save parties significant time and expense by avoiding the traditional "linear" review of potentially-relevant documents. When predictive coding is utilized, lawyers typically manually review and code sample batches of documents known as "seed sets" to train the computer program how to recognize relevant documents. While some have disputed the accuracy of computer assisted review, a 2011 study published in the Richmond Journal of Law and Technology concluded that computers more accurately coded documents than their human counterparts. See Maura R. Grossman & Gordon V. Cormack, Technology-Assisted Review in E-Discovery Can Be More Effective and More Efficient Than Exhaustive Manual Review, XVII RICH. J.L. & TECH. 11, 5 (2011) (arguing that "the levels of performance achieved by two technology-assisted processes exceed those that would have been achieved by . . . law students and lawyers employed by professional document-review companies – had they conducted a manual review of the entire document collection").
However, it must be noted that predictive coding is only appropriate for cases involving a large number of documents. Indeed, Vice Chancellor Laster recognized that the type of indemnification claim at issue in EOHB "can generate a huge amount of documents." (The transcript of Vice Chancellor Laster's bench ruling can be found here). Predictive coding has slowly gained acceptance in other jurisdictions besides Delaware. For example, in the United States District Court for the Southern District of New York, Magistrate Judge Andrew Peck recently endorsed its use in Monique Da Silva v. Publicis Groupe & MSL Group, Case No. 11-cv-01279 (S.D.N.Y. Feb. 24, 2012).
We expect that computer assisted review will continue to gain wider acceptance, and soon will become a standard process for the review of electronically stored information in disputes involving a large number of documents.