The early stages of large litigations have long been consumed by manual document review—namely—junior lawyers who spend weeks or months culling through mountains of documents at great expense to the client. Fortunately, such expenses could decrease substantially under a recent decision.1 Last week, a federal judge in the Southern District of New York approved the use of computer-assisted review to streamline the document review process, potentially cutting hundreds of thousands of dollars in costs from large litigations.

The decision, Moore v. Publicis Groupe & MSL Group, approved the use of predictive computer-assisted coding software to review millions of electronic documents in discovery. Predictive coding uses sophisticated algorithms to determine documents' relevance in conjunction with human review. The process involves reviewing and coding a "seed set" of documents by attorneys. The software then identifies properties of those documents to code other documents. "As the senior reviewer continues to code more sample documents, the computer predicts the reviewer's coding (or the computer codes some documents and asks the senior reviewer for feedback)," Magistrate Judge Peck wrote. "When the system's predictions and the reviewer's coding sufficiently coincide, the system has learned to make confident predictions for the remaining documents." Magistrate Judge Peck estimated that attorneys would only have to review a few thousand documents to effectively "train" the computer.

The Moore plaintiffs objected to this approach, claiming it failed to meet reliability standards and otherwise violated the Federal Rules of Civil Procedure. According to Magistrate Judge Peck, however, "computer-assisted review 'works better than most of the alternatives, if not all of the present alternatives'" without the cost. In adopting Magistrate Judge Peck's ruling, Judge Carter recognized that "ESI protocol contains standards for measuring the reliability of the process and the protocol builds in levels of participation by plaintiffs." He explained that "search methods [in Moore] will be carefully crafted and tested for quality assurance," because the plaintiffs may provide keywords and review both the documents and the "issue coding" prior to production. Judge Carter also recognized that parties may raise issues of relevance to the court prior to final production of documents and may challenge, once again, whether the use of protective coding software is the best search method available. He concluded that plaintiffs' characterization of predictive coding as unreliable was "speculative."

Judge Carter also recognized that traditional means of manual review with keyword searches is often flawed. He noted that "such review is prone to human error and marred with inconsistencies from the various attorneys' determination of whether a document is responsive."

This is an extremely important decision, particularly for defendants in large litigations. In essence, the Moore decision can be used as a road map to significantly diminish costs.