On Friday, February 24, New York Magistrate Judge Andrew Peck issued an opinion and order in Da Silva Moore v. Publicis Groupe & MSL Group, 11 Civ. 1279 (ALC) (AJP) (S.D.N.Y. Feb. 24, 2012), the first documented case to recognize predictive coding as an acceptable method of searching for electronically stored information ("ESI"). "Predictive coding," or computer-assisted review, is a process that can replace some human review with computer review, thereby reducing the cost of large-scale, e-discovery reviews. A typical technique is to identify a suitable "seed set" of documents that are reviewed by case attorneys and coded for relevancy, privilege, and other factors. Based on that coding, the system is trained to learn what is relevant and what is not; then the system itself predicts or suggests documents that are potentially relevant or similar across the document collection.
The Da Silva Moore opinion may go a long way toward allaying fears of lawyers about the use of this technology because they "no longer have to worry about being the 'first' or 'guinea pig' for judicial acceptance of computer-assisted review." Id. at 25. However, the opinion notes that the use of predictive coding is not suitable for all cases. Clients and their counsel must determine on a case-by-case basis whether predictive coding is appropriate for the case at hand.
The opinion provides insight for clients and counsel when deciding whether or not to adopt the use of predictive coding technology. Judge Peck indicated a greater willingness to accept the parties' ESI search protocol because both sides agreed to its use (at least in concept), and because of defendants' willingness to follow a "transparent" process. Id. at 22. In Da Silva Moore, the defendants agreed to provide plaintiffs with all non-privileged documents used to "seed" the predictive coding system (even the documents judged not relevant to the case), so that plaintiffs could review defendants' relevance coding being used to train the computer.
In addition, Judge Peck noted that fine-tuning the computer predictions is an iterative process, which typically requires repeated sampling and adjustment until the computer algorithms reach acceptable levels of accuracy. Judge Peck stated that it is "unlikely that courts will be able to determine or approve a party's proposal as to when review and production can stop until the computer-assisted review software has been trained and the results are quality control verified." Id. at 23.
Judge Peck further opined that, once the computer-generated results are proved to be at least as reliable as human review results, at least in identifying critical documents, it is not necessary to understand the computer algorithm used to achieve those results. Id. at 4. This calls to mind the old expression that "the proof of the pudding is in the eating." While that is a useful test for judging the efficacy of predictive coding, it should be noted that the test has yet to be passed in the Da Silva Moore case or any other publicly reported litigation. Further decisions in this case or others may illuminate how well the software ultimately performs.
In summary, Judge Peck's opinion provides a useful precedent supporting attempts to use this potentially cost-saving technology, but also identifies some of the associated risks and complexities. Careful planning, cooperation between counsel, and high-quality review of a sufficient sample of seed documents can be critical components of success. Companies facing large-scale document reviews should consider the potential use of computer-assisted review technology to improve speed and reduce costs. E-discovery counsel who already have experience using such technology can provide valuable guidance in this regard.