Predictive Coding Upheld by District Court: Judge Carter Endorses Judge Peck’s Approval of Computer-Assisted ESI Review

On March 2, 2012, we reported on Magistrate Judge Andrew Peck’s February 24, 2012 decision in Monique Da Silva Moore, et al., v. Publicis Groupe & MSL Group, Civ. No. 11-1279 (ALC)(AJP) (S.D.N.Y. Feb. 24, 2012), wherein Judge Peck issued the first judicial opinion approving the use of predictive coding “in appropriate cases.”  On April 25, 2012, District Judge Andrew L. Carter, Jr. rejected plaintiffs’ bid to overturn that decision, and cleared the way for the use of computer-assisted ESI review in this case and others. Monique Da Silva Moore, et al., v. Publicis Groupe & MSL Group, Civ. No. 11-1279 (ALC)(AJP) (S.D.N.Y. Apr. 25, 2012).

Judge Peck originally approved the use of predictive coding in Da Silva Moore based on several factors, including the parties’ theoretical agreement with the concept, the need to review an enormous data set in excess of 3 million documents, the superiority of computer-assisted review over the alternatives (manual review or keyword searches), the need for cost effectiveness and proportionality under Fed. R. Civ. P.26(b)(2)(C) and the “transparent” process defendants proposed. But he did so over plaintiffs’ continued objections, which were detailed in the papers plaintiffs submitted to District Judge Carter on February 22, 2012, and which Judge Peck essentially disposed of in his opinion issued two days later.

When defendants responded to plaintiffs’ objections on March 7, 2012, plaintiffs’ cried foul and requested leave of District Judge Carter to respond. In sum, plaintiffs contended Judge Peck’s written analysis went well-beyond the rationale he articulated from the bench at the February 8, 2012 hearing on which his opinion was based and, further, relied on materials not previously discussed or referenced by the parties. Plaintiffs also objected to the inequity of allowing defendants to respond to plaintiffs’ objections after having the advantage of digesting Judge Peck’s decision. District Judge Carter granted plaintiffs’ request, and plaintiffs filed further detailed objections to defendants’ protocol and Judge Peck’s rulings. For good measure, plaintiffs also informally asked Judge Peck (by letter) to recuse himself from the case based largely on his participation and comments at e-discovery conferences discussing his support for the use of predictive coding. Judge Peck refused and, in his April 2, 2012 Order, defended himself against plaintiffs’ accusations of bias and specifically cautioned plaintiffs to “re-think their scorched earth approach” to the litigation. Judge Peck also predicted his admonition would fall on deaf ears, and it did; on April 13, 2012, plaintiffs formally moved for his recusal, which is pending.

Against this background, District Judge Carter clearly had much to contend with in reviewing and considering plaintiffs’ objections. The Court began by summarizing plaintiffs’ arguments, to the effect that Judge Peck’s decision was erroneous and contrary to law because the ESI protocol lacks “generally accepted reliability standards” and, as such, its use violates Fed. R. Civ. P. 26 and Fed R. Evid. 702, and that Judge Peck failed to hold an evidentiary hearing and therefore adopted the protocol on an insufficient record and further went beyond the record by considering documents not discussed in the parties initial submissions. Although Judge Carter next proceeded to set forth the case law establishing the applicable “highly deferential standard of review”, the Court might just as well have started or even abbreviated its discussion with the concluding sentence of its opinion: “[t]he Court reminds the parties that it affords Judge Peck’s non-dispositive rulings great deference and that magistrate judges generally have broad latitude with respect to discovery issues.”

Judge Carter adopted Judge Peck’s rulings, characterizing them as well-reasoned and generally praising their consideration of the “potential advantages and pitfalls” of predictive coding software. Rejecting plaintiffs’ objections, the Court specifically noted that the protocol contained “standards for measuring the reliability of the process” and that “the protocol builds in levels of participation by Plaintiffs.” The Court highlighted plaintiffs’ ability to raise concerns about the relevance of culled documents with Judge Peck before final production and found “insufficient evidence” to conclude that the parties’ use of predictive coding software would deprive plaintiffs of “liberal discovery.”

The Court also found plaintiffs’ reliability arguments “premature” and rejected their arguments regarding the need for an evidentiary hearing, explaining that “if the method appears unreliable as the litigation continues and the parties continue to dispute its effectiveness,” Judge Peck could hold the evidentiary hearing later to address those issues. Judge Carter concluded the point by noting that “if the method provided in the protocol does not work or if the sample size is indeed too small to properly apply the technology, the Court will not preclude Plaintiffs from receiving relevant information, but to call the method unreliable at this stage is speculative.” In concluding his review, Judge Carter observed “[t]here simply is no review tool that guarantees perfection” and specifically identified the risks inherent in manual and keyword searching. Ultimately finding no basis to disturb Judge Peck’s findings that the use of predictive coding software was “more appropriate than keyword searching” in “this particular case,” Judge Carter upheld Judge Peck’s rulings and denied plaintiffs’ objections.

In our initial blog posting discussing Judge Peck’s decision, we indicated that Da Silva Moore is significant because it opened the door for parties to consider the use of computer assisted review in appropriate cases without the fear that a party or judge will resist based on a perceived lack of acceptance. Judge Carter’s adoption of Judge Peck’s rulings constitutes a big, bold underlining of that statement, and regardless of the outcome of any further proceedings, we can expect a great deal of interest in and use of predictive coding technology going forward.