Latest Thinking | eDiscoverings

Court: "Counsel no longer have to worry about being the ‘first’ or ‘guinea pig’ for judicial acceptance of computer-assisted review”

In Da Silva Moore v. Publicis Groupe, et al., No. 11 Civ. 1279 (ALC) (AJP), 2012 WL 607412 (S.D.N.Y. Feb. 24, 2012 ), a court for the first time endorsed the use of predictive coding as an acceptable way to search for relevant ESI.  Although the district court decision has minimal precedential value, particularly since both parties had agreed to the use of predictive coding, the court’s analysis suggests that more courts in the future are likely to endorse its use even over the objection of one of the parties.

The predictive coding protocol approved by the court involves a seed set of approximately 2,399 documents (out of more than 3 million total e-mails) that senior attorneys will review and use to train the predictive coding software.  The defendant agreed to share the seed set with plaintiffs to confirm that plaintiffs agreed with defendant’s relevance decisions so as to ensure that the computer is being trained properly.  A minimum of seven additional rounds of review of at least 500 documents from different concepts will follow and be used to further train the computer.  These additional documents also will be shared with plaintiffs.  Additional rounds of review will take place if necessary to stabilize the system.  Defendant agreed to review for privilege all of the relevant documents identified by the computer, but reserved the right to seek relief from the court if more than 40,000 documents are identified as relevant.

The court’s decision is welcome news for litigants.  In today’s litigations, which often involve millions of potentially responsive documents, “linear manual review is simply too expensive….”  2012 WL 607412, at *9.  Thus, the discovery paradigm must change if there is any realistic chance of reducing the costs of electronic discovery.  Innovative solutions, such as the use of predictive coding, will go a long way towards helping to achieve that goal.  Indeed, even though the court recognized that “computer-assisted review is not perfect, the Federal Rules of Civil Procedure do not require perfection.”  Id. at *11.  In other words, there simply is no reason to search for, review and produce every conceivable document that might be potentially relevant in a dispute.  Discovery should be about finding the most relevant documents, not every possible responsive document.  As the court explained, “[c]ourts and litigants must be cognizant of the aim of Rule 1, to ‘secure the just, speedy , and inexpensive determination’ of lawsuits…[as] further reinforced by the proportionality doctrine set forth in Rule 26(b)(2)(C).”  Id.

Please do not hesitate to contact David Lender at or (212) 310-8153 if you would like to discuss this case or any other electronic discovery issues.