On April 26, 2012, Judge Andrew Carter (S.D.N.Y.) upheld the groundbreaking decision of Magistrate Judge Andrew Peck in Da Silva Moore v. Publicis Groupe, No. 11-civ-1279 (ALC), 2012 WL 1446534 (S.D.N.Y. April 26, 2012) (Carter, J.). That decision had approved, for the first time, the use of predictive coding software to review and produce ESI. Da Silva Moore v. Publicis Groupe, No. 11-civ-1279 (ALC) (AJP), 2012 WL 607412 (S.D.N.Y. Feb. 24, 2012) (Peck, M., J.). The affirmance provides further judicial acceptance of computer-assisted review as an alternative to costly and error-ridden linear human review of large volumes of ESI. This article briefly discusses Judge Carter’s affirmance, the import of judicial acceptance of predictive coding for corporate counsel, and the future fate of predictive coding.
Judge Carter’s Decision
Judge Carter reviewed Judge Peck’s decision under Fed. R. Civ. P. 72(a)’s deferential standard for discovery orders. He held that Judge Peck’s allowance of “predictive coding” was within the reasonable exercise of the court’s discretion: “[T]he Court adopts Judge Peck’s rulings because they are well reasoned and they consider the potential advantages and pitfalls of the predictive coding software.” Da Silva Moore, 2012 WL 1446534 at *2 (Carter, J.).
Judge Andrews thoroughly reviewed the ESI Protocol approved by Judge Peck. Id at *2. He referenced the protocol’s protections of plaintiff, including standards for measuring reliability, quality control and the opportunity for plaintiff to challenge the end results. Id. He relied heavily upon the transparency of, and plaintiff’s counsel’s active participation in, the process employed by defendant. Id. Plaintiff’s counsel had input into the selection of key words used to generate the “seed sets” of documents reviewed by defendant to “train” the software, access to the documents themselves, and defendant’s coding of them. Finally, plaintiff could dispute defendant’s coding decisions. 2012 WL 607412 at *5 (Peck, M.J.).
Judge Carter also agreed with Judge Peck’s assessment of the accuracy of human review versus predictive coding: “[i]t is difficult to ascertain that the predictive software is less reliable than the traditional keyword search.” Compare Id. at *9 with 2012 WL 1446534 at *3 (Carter, J.). Judge Carter offered this practical observation:
There simply is no review tool that guarantees perfection. The parties and Judge Peck have acknowledged that there are risks inherent in any method of reviewing electronic documents. Manual review with keyword searches is costly, though appropriate in certain situations. However, even if all parties here were willing to entertain the notion of manually reviewing the documents, such review is prone to human error and marred with inconsistencies from the various attorneys’ determination of whether a document is responsive.
Id.
The Import Of Judicial Acceptance Of Predictive Coding
Corporate counsel see judicial acceptance of predictive coding as an important step in battling the ever-increasing costs of e-discovery, costs that threaten the viability of the present civil justice system. Tim Pratt, General Counsel of Boston Scientific and Secretary-Treasurer of the Federation of Defense & Corporate Counsel, commented on the decision:
Judge Carter's opinion accomplishes something we all should want – a better, quicker, cheaper way to get to the truly relevant documents in litigation. The first rule in the Federal Rules of Civil Procedure lays out the standard. The rules should be construed and administered to secure the just, speedy and inexpensive determination of every action and proceeding. With the endless proliferation of electronically stored information within companies, there should be general alignment on the importance of coming up with a reliable method of identifying the documents that count. Predictive coding is one good way to do that. There will likely be others.
Boston Scientific has used predictive coding and seen demonstrable reductions in e-discovery costs.
These thoughts were echoed by Marc Polk, Associate General Counsel/Litigation for Covidien, who noted that “[i]n the right context, predictive coding can be a valuable tool to manage costs and advance what should be the goal of everyone involved – identifying and obtaining information that is truly responsive and relevant to the matter.”
Apart from cost savings, computer-assisted document review can produce increased control and predictability. The process is decidedly different from human linear review, where an army of junior attorneys, paralegals, and/or contract attorneys review documents in cases about which they know little or nothing. See 2012 WL 607412, at *2 (Peck, M.J.). In predictive coding, seed sets are reviewed by a small group of counsel with detailed command of the facts and issues. Predictive coding software takes the work product from these senior reviewers and “trains” itself to identify responsive documents. The same small group of senior reviewers engages in iterative reviews of predictive coding results, refining the responses along the way. See id.
Will Courts Continue To Embrace Predictive Coding?
It is important to keep the Da Silva decisions in perspective. Both are preliminary discovery rulings that on their face reserve final judgment on the ultimate reliability of this new approach. Plaintiff can still challenge the end result of the computer-assisted production. But the opinions strike recurrent themes and demonstrate a judicial willingness to look for solutions to the problems inherent in reviewing large volumes of ESI.
From Marc Polk’s perspective, predictive coding is not that huge of a leap from tools already in use:
Predictive coding is really a natural progression in the evolution of tools to meet the demands of discovery of electronic information more efficiently. It can help prevent fishing expeditions and reign in overly burdensome discovery demands. It takes term searching and coding to the next level to not only help identify and prioritize relevant information more accurately, but also to more quickly eliminate non-responsive material.
Whether other courts follow the decisions in Da Silva may depend on how effectively counsel demonstrate to them the real limitations of human linear review. Thinking of human review as the “gold standard” is “a myth.” 2012 W.L. 607412 at *9 (Peck, M.J.). A growing body of data supports the accuracy of predictive coding, particularly when measured against the mythical gold standard. See id. Counsel will likely insist that any party objecting to the use of predictive coding delineate an alternative proposal, its cost, and its outcome.
The federal court’s recognition of the need for cooperation, disclosure, and transparency amongst counsel in the e-discovery arena also may weigh in favor of future acceptance of predictive coding. The potential cost savings of predictive coding will motivate producing parties to embrace disclosure and transparency. The Da Silva ESI Protocol provides a roadmap that would make it difficult for objecting parties to complain.
Predictive coding is only a tool, “not a magic, Staples-Easy-Button, solution, appropriate for all cases.” Id. at *8. But it is a tool that offers great promise to ameliorate issues that burden our civil justice system. Boston Scientific’s Tim Pratt looks to the future with some optimism:
Corporate counsel want to control the ever-escalating costs of e-discovery. They also want predictability and manageability. One can achieve all of that when open-minded litigants approach the process in an innovative, problem-solving way. Judges like Magistrate Judge Peck and District Court Judge Carter are dealing with the issues in a way that accomplishes the goals of the Federal Rules of Civil Procedure and that requires continued collaboration and cooperation by litigants. That, together with the efforts of organizations like Lawyers for Civil Justice in championing changes that level the playing field and allow for a truly just, speedy and inexpensive resolution to every dispute, provides some hope in managing, if not resolving these serious issues.
Published May 17, 2012.
