The Brave New World Of Predictive Coding

Monday, August 1, 2011 - 01:00

The Editor interviews Alitia Faccone , Partner in the Business and Financial Services group of McCarter & English, LLP.

Editor: In its July interview, Recommind states that it has developed patented technology that achieves high levels of savings in e-discovery and in other applications. What are your general reactions?

Faccone: The development of a program such as Recommind's is extremely important, but it does not represent a complete sea change. Emerging technologies such as concept clustering and predictive coding are not entirely new to the marketplace, and have long been in development and increasing use. What is exciting is the increased levels of awareness and acceptance of these tools. Predictive coding represents a major step forward in streamlining the e-discovery process, but the software must "learn" the coding from an initial round of manual attorney coding. Recommind's technology, and tools like it, present real opportunities for efficiency and cost savings, but clients must recognize that as a result of document review, production has not been reduced to simply hitting a switch. Any document review tool is just that: an implement that is only as effective as the attorneys who use it. Moreover, attorneys must continue to oversee the coding process by conducting quality control checks and adjusting the coding for any errors. In that sense, it is much like the initial transition from paper-based to computer-based document review.

Editor: What implications do that and similar approaches have on the need to address ESI issues in the FRCP?

Faccone: Rule changes should only be made after a thorough, deliberative and analytical process. For example, the Advisory Committee on Civil Rules is currently considering whether rule changes to address preservation and related sanctions issues for a failure to preserve are viable and desirable. But even these deliberations have come after much dialogue by leading practitioners and courts who have been considering the issue for some time, such as at last year's Duke Conference on Civil Procedure and eDiscovery.

In amending the FRCP to reflect the advent of e-discovery in 2006, the drafters had to balance two competing objectives. On the one hand, in terms of preservation, collection and review of information, electronic documents are fundamentally different from paper documents such that some updates to the rules were necessary. On the other hand, the new rules were not made too specific to the current state of the art, because technology changes so frequently. The latter consideration is the reason that, for instance, the initial e-discovery amendments did not include a suggested specification in Rule 26(b)(2)(B) that backup tapes are per se inaccessible. Balancing these considerations is a formidable task and one that must be kept in mind in any discussion of potential rule amendments. So, while predictive coding bears on a number of procedural issues, because the use of this and similar technology remains relatively novel, changes to the rules are not necessarily warranted or advised.

However, the better question may be how technology is impacting the implementation of the federal rules. Take Federal Rule of Evidence 502. When documents are produced as a result of automated coding and are not individually reviewed before they are produced, it is possible that some sort of privilege issue will arise i.e., privileged documents will be produced or non-privileged documents will be held back. As Magistrate Judge Paul Grimm pointed out in a recent law review article, FRE 502 was intended to provide a mechanism to reduce both the anxiety and costs associated with privilege review by enabling lawyers and their clients to utilize technology to reduce the amount of painstaking, manual document-by-document review or by entering into agreements to protect parties against waiver. But the rule has yet to fulfill its purpose because parties have either overlooked its benefits or are not entirely comfortable relying on the rule as it has not been interpreted consistently, preventing parties from comfortably predicting its interpretation in court.

This point is best illustrated by the recent cases of Datel Holdings, Ltd. v. Microsoft Corp., 2011 WL 866993 (N.D. Cal. Mar. 11, 2011) and Mt. Hawley Ins. Co. v. Felman Prod., Inc. , 2010 WL 1990555 (S.D. W. Va. May 18, 2010). In Datel the court held that a software glitch resulting in mistaken production of documents should be deemed inadvertent and the privilege had not been waived. However, in Mt. Hawley , the court concluded that the producing party had failed to perform "critical quality control sampling to determine whether their production was appropriate and neither over inclusive nor under inclusive" and that the inadvertent disclosures were not caused solely by the failure of the Concordance database to completely index the documents, as the plaintiff had argued. Although the factual circumstances of each of these cases were not directly analogous, the divergent results demonstrate the inconsistency in interpretation and, to a certain degree, justify a reluctance to take full advantage of the benefits of technology and rules that have been implemented that recognize its benefits.

Additionally, the circuits are far from consensus on the imposition of e-discovery sanctions. While this is as much a matter of case law as of the rules themselves, one of the major unanswered questions is how the use of predictive coding will affect the imposition of sanctions when coding causes impropriety in discovery.

Editor: How do you believe the courts will react?

Faccone: In 2011, the word on the lips of judges in the know is proportionality . Last October, The Sedona Conference published its commentary on proportionality, bringing heightened awareness to the concept. Considering the proliferation of data and its impact on litigation, courts are beginning to view predictive coding as a positive development and understand its value relative to proportionality in litigation. In accordance with FRCP 26(b)(2)(C), a court, in reply to a motion or of its own volition, may act to limit the scope of discovery so that discovery is not unduly burdensome. To that end, predictive coding might be used to take the measure of potentially discoverable information at a relatively early state of litigation. This would promote the kind of efficiency and cost savings courts (and, indeed, litigants) favor.

Few, if any actual, rulings involving predictive coding have been issued, but that is not surprising considering how new the technology is. But in more informal settings, certain influential judges have expressed an attitude of guarded optimism. For example, in a recent podcast, Judge Grimm noted that, because technology "will always outpace us," someone must have the courage to go first.

But no matter how many courts embrace predictive coding, attorneys and their clients must remember that such technology has to be used skillfully. Courts have not hesitated to reprimand attorneys who abdicate their role in the e-discovery process, as when custodians are allowed to pick and choose responsive documents ("custodial self-selection") or when attorneys fail to verse themselves in the technical aspects of discovery collection and production. In a recent case, Roffe v. Eagle Rock Energy , in Delaware, Vice Chancellor Laster explicitly stated, "[Y]ou do not rely on a defendant to search their own e-mail systemThere needs to be a lawyer who goes and makes sure the collection is done properlywe don't rely on people who are defendants to decide what documents are responsive." Courts are now more technically savvy and expect the lawyers and parties to demonstrate more technical knowledge and are less forgiving when they don't.

Case in point - recent changes to the New York rules of practice may be a harbinger of things to come. One year ago (August 18, 2010), the New York Trial Court Uniform Rules 202.12(b) and 202.70(g) were amended to require that attorneys must be "sufficiently versed" on the client's technological systems and able to speak "competently" to all e-discovery issues at preliminary conferences. Counsel may bring technical experts to assist with e-discovery discussions at the conference, but experts are no substitute for attorney knowledge. These new rules have teeth: noncompliance is punishable and may lead to a finding of default judgment.

Editor: To what extent do advances in e-discovery technology help corporate legal departments and law firms become proactive in their data management and litigation readiness?

Faccone: If there is one point we stress to clients, it is that a proactive approach to data management is crucial - both within and outside the litigation environment. There is nothing novel about the necessity to be prepared, but in today's world of e-discovery that advice has become more sound and more crucial. When you have in place a system under which documents are routinely categorized and filed - and, just as important, destroyed if they need not be kept - you will avoid the high price of having your legal team retroactively sort those documents in response to a document request. The sooner attorneys can move from "sorting" to "reviewing," the more efficient the process will be. By the same token, when documents are already organized, less predictive coding will be necessary just to make sense of the documents under consideration.

Clients also have the option of utilizing other developing technology such as "litigation hold" software that can be implemented at the moment of a lawsuit's initiation to retain relevant documents with greater ease and accuracy. The question of adequate preservation efforts is the source of numerous e-discovery disputes and, as a result, has led to the imposition of sanctions in many cases. Many of these disputes have turned on the moment at which litigation may be "reasonably anticipated," a question for which a clear-cut answer has proven elusive - precisely why an amendment to the FRCP is being explored. In the face of this uncertainty, clients should be all the more ready to spring into action at the moment preservation is required. Preparation is the key, and technology is making preparation and implementation easier and far more cost-effective.

Editor: Will e-discovery technology ever be able to replace human review of large numbers of documents for relevance and privilege? Does this analysis change depending on the kind of matter being handled?

Faccone: E-discovery technology must continue to complement, not replace, the work of human reviewers. The ethical rules that all lawyers must follow require that lawyers not abdicate their responsibilities to a computer. What will continue to change is the recipe for human and technological review in e-discovery.

This past March, a New York Times article created a stir by suggesting that data review software would one day replace document review by lawyers. Although it is true that software - and predictive coding in particular - is likely to reduce the number and way in which contract attorneys are deployed for document reviews, it is equally true that software presents new and different opportunities for regular practitioners. The future is not about the decline of the attorney but about the rise of the legal/IT attorney, i.e., an attorney who can combine knowledge of the technology and the law to deliver the best possible discovery results to a range of clients. At our firm, cultivating these "attorneys of the future" is a conscious effort. As I mentioned earlier, we make it our business to ensure that attorneys at all levels are up to speed on the latest developments and understand the intersection of the mechanics and the content.

Editor: Given the rapid development of e-discovery and related technology, what steps is your firm taking to assure that your clients are taking advantage of the most effective and economical cutting-edge technologies with respect to various aspects of EDRM and the stages of ESI collection, preservation and production?

Faccone: Our firm, our e-discovery lawyers in particular, make it an absolute priority to stay on top of the latest developments in e-discovery technology. The proliferation of vendors and types of technology solutions provide many options for our lawyers to recommend based on the specific needs and resources of a given client, understanding which options are viable and can best suit our clients' needs. We believe it is our responsibility today for our lawyers and legal IT professionals to continue to evaluate these emerging technology solutions in order to make informed and strategically appropriate recommendations. As a result, our understanding of how to select and use such alternative review methodologies is an integral component of the value that we bring to any given matter. These efforts are firmwide: from senior partners to new associates. The end result is that when a client walks through the door, we already possess the knowledge to lay out all the options and determine the best way to move forward. We try to achieve a balance between the quality of service a vendor provides and the final cost to the client. We want clients to see technology as part of an overall portfolio of services, not as a mere cost-cutting measure.

Please email the interviewee at with questions about this interview.