Predictive coding, the use of computer algorithms and machine learning as part of document review, has been billed as the next generation of technology in e-discovery. For years, e-discovery service providers — along with some jurists and e-discovery industry veterans — have sung its praises as, at minimum, a complement to the standard document review process, and possibly as a replacement for it.
In reality, there is no one-size-fits-all solution for managing e-discovery. In many situations, the use of advanced technology, perhaps even predictive coding, may enhance the speed, efficiency, and quality of review of electronically stored information (ESI). In many others, it will not. Questions loom as to when and how best to leverage advanced technology and, with respect to predictive coding, its cost, effectiveness, and level of acceptance by parties and judges.
Magistrate Judge Andrew Peck of the Southern District of New York, well-known as someone who not only accepts predictive coding, but also actively promotes its use, has become synonymous with the topic of predictive coding. His advocacy through articles, panel appearances, and decisions such as Da Silva Moore v. Publicis Groupe, 287 F.R.D. 182 (S.D.N.Y. 2012) and Rio Tinto PLC v. Vale S.A., 306 F.R.D. 125 (S.D.N.Y. 2015), has helped to convert many in the e-discovery industry from skeptics to believers with respect to the use of predictive coding. A true believer himself, Judge Peck has often urged parties to use predictive coding, which he calls “technology-assisted review” or TAR, whenever appropriate.
Would Judge Peck’s admiration and advocacy for predictive coding lead him, upon a request by the opposing party, to force a responding party to use it against that party’s own wishes? Judge Peck recently faced this issue, putting potential use of predictive coding at odds with established precedent and procedure regarding how to conduct discovery.
Hyles v. New York City
In Hyles v. New York City, 2016 WL 4077114 (S.D.N.Y. Aug. 1, 2016), the plaintiff sued the City of New York (the City), asserting claims of discrimination and a hostile work environment. The case was referred to Judge Peck due to significant disputes regarding discovery.
Judge Peck had ruled that the parties should stage discovery — an effective tool to help manage the discovery process and to promote proportionality in the scope of discovery. As part of staged discovery, after the City produced information from nine custodians, if the plaintiff could demonstrate that six other custodians “had relevant, unique and proportional ESI,” id. at 1, Judge Peck would consider having the City search them too.
What was at issue was how the City would search the ESI gathered from those nine custodians. The plaintiff wanted the City to use predictive coding/TAR, claiming it was a “more cost effective and efficient method of obtaining ESI from Defendants.” Id. at 2 (citations omitted). The City, however, “declined, both because of cost and concerns that the parties, based on their history of scope negotiations, would not be able to collaborate to develop the seed set for a TAR process.” Id. (citation omitted).
Before Judge Peck was the question whether the plaintiff could force the City to use predictive coding, when the City preferred to use keyword searching. As succinctly and effectively put by Judge Peck, “[t]he short answer is a decisive ‘NO.’” Id. at 1.
Judge Peck’s decision rested on some critical findings and observations about modern e-discovery practice. First, he relied on The Sedona Conference, which has been the leading think tank on e-discovery for nearly two decades. He noted Principle 6 of The Sedona Principles: Second Edition, which states: “Responding parties are best situated to evaluate the procedures, methodologies, and technologies appropriate for preserving and producing their own electronically stored information.” Id. at 3 (citation omitted). This principle is powerful, especially in today’s e-discovery climate in which responding parties are preserving, collecting, reviewing, and producing massive volumes of data, often from complex systems. Judge Peck forcefully and clearly ruled on this topic, stating that “the City as the responding party is best situated to decide how to search for and produce ESI responsive to Hyles’ document requests.” Id.
Second, while some, including Judge Peck, are convinced that the technology used in predictive coding is “cheaper, more efficient and superior to keyword searching,” id. at 2 (citation omitted), and such advanced technologies have become generally accepted as a permissible means of conducting e-discovery, it is still a process to be used in appropriate situations, as one of many available tools for document review. And, as the City asserted, the potential for the parties to become embroiled in satellite disputes about the TAR process itself might very well make the process more, not less, costly. As a result, while TAR could potentially be helpful, its use is not required. As Judge Peck wrote: “There may come a time when TAR is so widely used that it might be unreasonable for a party to decline to use TAR. We are not there yet.” Id. at 3.
Third, while it is critical for parties to cooperate as part of the e-discovery process, cooperation may not be used as a sword. As Judge Peck writes:
I am a signatory to and strong supporter of the Sedona Conference Cooperation Proclamation, and I believe that parties should cooperate in discovery … . The December 1, 2015 Advisory Committee Notes to amended Fed. R. Civ. P. 1 emphasized the need for cooperation. Cooperation principles, however, do not give the requesting party, or the Court, the power to force cooperation or to force the responding party to use TAR.
Id. at 2 (internal citations omitted).
Fourth, perfection is not required as part of the document review and production process. As Judge Peck notes: While Hyles may well be correct that production using keywords may not be as complete as it would be if TAR were used, the standard is not perfection, or using the ‘best’ tool, but whether the search results are reasonable and proportional.” Id. at 3 (internal citations omitted).
The emergence in the document review process of advanced technology tools that seek to promote accuracy and speed and to reduce costs has the potential to be the next frontier of e-discovery, but, as Judge Peck notes in Hyles, we are not there yet.
While parties should always focus on improving quality, efficiency, and defensibility in e-discovery and consider what technologies and processes are best suited to the document review process, the analysis of and decision whether to utilize technology such as predictive coding should be made on a case-by-case basis. As Hyles demonstrates, many parties are rightfully concerned about the additional cost and potential for satellite disputes associated with agreeing on a predictive coding protocol — costs and disagreements that are not as prevalent when parties employ more traditional document review tools. Many parties continue to believe that any perceived savings derived from using predictive coding may be swallowed by the additional lawyer time and motion practice that often results when parties cannot agree on the predictive coding process.
Judge Peck, a self-described “judicial advocate for the use of TAR in appropriate cases,” is also a jurist with deep experience in the complexity of modern e-discovery, and, as shown in Hyles, understands that “it is not up to the Court, or the requesting party[,]” to force a responding party to use a specific technology or process as part of document review. Id. As such, in addition to providing a pragmatic perspective on modern e-discovery practice, Hyles is an important decision that reaffirms the long-held notion that responding parties are best situated to determine how to respond to discovery requests.
H. Christopher Boehning and Daniel J. Toal are litigation partners at Paul, Weiss, Rifkind, Wharton & Garrison. Ross M. Gotler, e-discovery counsel, and Lidia M. Kekis, e-discovery attorney, assisted in the preparation of this article. This article also appeared in the New York Law Journal, an ALM sibling of Cybersecurity Law & Strategy.
The views expressed in the article are those of the authors and not necessarily the views of their clients or other attorneys in their firm.