H. Christopher Boehning and Daniel J. Toal ()
Technology-assisted review, or TAR, is undoubtedly gaining traction in e-discovery practice in complex civil litigations and regulatory investigations. However, practitioners and judges still grapple with inconsistencies and unresolved issues regarding its use and applicability, in no small part due to a shortage of legal opinions on the topic and a lack of consistency in the decisions that do exist.
In an effort to restate the current law on TAR, The Sedona Conference, a leading think tank on ediscovery law and practice, recently released its TAR Case Law Primer. This Primer does not propose TAR best practices or endorse specific TAR methodologies; instead, it “analyzes decisions from those courts that have been required to opine on the efficacy of TAR in a variety of circumstances and explores the evolution in the courts’ thinking[.]” While TAR can be used as an umbrella term to describe many types of advanced technology tools and processes used to aid in the document review process (such as email threading, concept searching, and automated clustering), the Primer uses it as a synonym for what is also called predictive coding. The Primer defines TAR as a “ process for prioritizing or coding a collection of Electronically Stored Information using a computerized system that harnesses human judgments of subject matter expert(s) on a smaller set of documents and then extrapolates those judgments to the remaining documents in the collection.”
In a recent decision, a court turned to the Primer for guidance in resolving a dispute that, ideally, the parties should have resolved on their own. The court, though, with an opportunity to move the law on TAR forward and bring some clarity on a key issue, may have muddied the waters even more.
In FCA US v. Cummins, 2017 WL 2806896 (E.D. Mich. March 28, 2017), the court was put in a difficult position when the parties asked it to intervene and resolve a discovery dispute. In devising their discovery protocol for electronic materials, the parties were at an impasse regarding pre-TAR culling, specifically “whether the universe of electronic material subject to TAR review should first be culled by the use of search terms.”
In a very short decision, the court “rather reluctantly” opined, despite believing that the parties should have been able to resolve the dispute on their own without judicial intervention. Without detailed explanation, the court simply stated, “having reviewed the letters and proposed orders together with some technical in-house assistance including a read of [the Primer], the court is satisfied that [the plaintiff] has the better position. Applying TAR to the universe of electronic material before any keyword search reduces the universe of electronic material is the preferred method. The TAR results can then be culled by the use of search terms or other methods.” While the decision did not go into detail in defining what it meant by TAR, specifically whether it was predictive coding or instead other types of advanced technology, it can be inferred from the court’s reliance on the Primer and from the reference to commonly-used predictive coding terms in the resulting discovery protocol order that the court was using the term “TAR” as a synonym for predictive coding.
The Primer includes a section on “Disputed Issues Regarding TAR,” explaining that “[a] number of decisions have addressed various disputed issues regarding the use of TAR. Many or all of these issues remain open, either because of a lack of consensus among the decisions, an absence of in-depth analysis in the decisions, the fact-specific nature of certain decisions, or the paucity of decisions addressing an issue.” One of the disputed issues covered is “Using Search-Term Culling Before TAR,” with four cases on the topic highlighted. While those cases may generally opine, in dicta, that, in a perfect world, TAR would be conducted prior to any keyword culling, if there is a common thread to the actual decisions, it is that, as noted in one of the decisions, “‘the standard for TAR is not perfection,’ nor ‘best practices,’ ‘but rather what is reasonable and proportional under the circumstances.’”
In other sections, the Primer reports on cases that generally come to the conclusion that even though there may be a preferred TAR best practice, all parties must do in the discovery process is meet their obligations under the Federal Rules of Civil Procedure, which require reasonable efforts in discovery. Such cases reference Principle 6 of the well-known Sedona Principles, which states, “Responding parties are best situated to evaluate the procedures, methodologies, and technologies appropriate for preserving and producing their own electronically stored information.”
Every case is different; there is no one-size-fits-all strategy regarding the specific TAR methodology to be used by responding parties. In some instances, the burden and expense of using TAR pre-culling on the entire document population may not be worthwhile in light of proportionality considerations. The potential high costs in processing and ingesting a large universe of data into a TAR process may persuade a party to consider utilizing pre-culling tools, such as keyword filtering, to target and remove likely non-responsive documents. As noted in an August 2016 report by the Coalition of Technology Resources for Lawyers, “The majority rule on this issue nonetheless remains that litigants can use search terms to remove non-responsive materials prior to running TAR on the remaining document population. This generally accords with notions of proportionality and reasonableness … [and,] [t]hus, courts will generally approve the combined use of TAR with other methodologies designed to reasonably reduce the size of the document population so long as the process results in productions of highly relevant information that are proportional under the circumstances.”
As such, a court that finds itself in a situation similar to the one in Cummins could, instead of choosing a particular TAR strategy as preferred, direct each responding party to do what it determines is reasonable and proportional and allow for further consideration if the receiving party can later show deficiencies in the process. As with many TAR-related decisions, Cummins should be taken in light of the particular circumstances of the case and not broadly applied to all future cases. While there may be isolated situations where a court may need to step in, as the court itself noted in Cummins, such matters are best resolved between the parties without judicial intervention.