In Monique Da Silva Moore, et al. v. Publicis Groupe & MSL Group, a decision from the United States District Court for the Southern District of New York explicitly recognized the use of predictive coding technology (i.e., computer-assisted review) as an appropriate method to satisfy a producing party’s review obligations. The decision by Magistrate Judge Andrew Peck holds that the use of predictive coding technology “is an acceptable way to search for relevant [electronically stored information] in appropriate cases.”
The decision will be widely heralded as approving the use of predictive coding. That conception oversimplifies the issue and misses the more important lessons of Judge Peck’s decision. Monique Da Silva Moore is instructive not because it holds that predictive coding is a now an acceptable technology. No such holding was necessary. Squire Sanders, like some other firms, already has been using predictive coding methods to meet its clients’ discovery obligations in a defensible and cost-effective manner. Rather, the decision highlights that the use of predictive coding – like other technologies and methods – is acceptable where it is employed in a manner that produces reliable and proportional results. As recognized by Judge Peck, the necessary inquiry by a court is not whether counsel has picked an appropriate technology; “it is the process used and the interaction of man and machine that the court needs to examine.”
In Monique Da Silva Moore, the court rejected as premature plaintiffs’ challenge to the reliability of the predictive coding review because Judge Peck concluded that the technology was being employed effectively under the circumstances: judgmental sampling decisions were made by senior attorneys; there was transparency regarding the document set used for training; certain non-corresponding “comparator” email sets were excluded from the scope of the coding; and documents deemed irrelevant would be sampled later to confirm the effectiveness of the training. The review method was upheld not because it included predictive coding technology, but rather because the technology was applied in an appropriate way likely to produce reliable results at a proportional cost.
Judge Peck also rejected the notion that the Federal Rules of Civil Procedure demand perfection from a review method. Monique Da Silva Moore cited well recognized empirical data establishing that other widely used review methods, including untested keywords and manual linear review, are far from perfect. Judge Peck observed that computer-assisted review, while also not perfect, was appropriate because when used appropriately it can produce higher recall (i.e., completeness of review results) and precision (i.e., accuracy of review results) at a proportional cost, particularly in cases with a large volume of data. This, not the particular technology at issue, is the real lesson of Monique Da Silva Moore.