More on the ethics of e-discovery: Predictive coding and other forms of computer-assisted review


john_barkettBy: John Barkett | Partner | Shook, Hardy & Bacon
Date: Wednesday, February 20, 2013

This paper is republished at with the permission of its author. John Barkett is a member of the ACEDS Advisory Board and will be presenting on computer-assisted review at the ACEDS 3rd Annual Conference and Exhibition.

Consider the following hypothetical.

Your client has been sued. Your case has been assigned to a district court judge and a magistrate judge who have both endorsed the Cooperation Proclamation published by the Sedona Conference. You had already met with your client and the client team to discuss search protocols and techniques with respect to electronically stored information to prepare your case as well as to prepare for production in anticipation of expected requests for electronically stored information. You contact the client’s general counsel to discuss the meet-and-confer session. You tell the general counsel that the issue of search protocols is going to come up. You expect your opponent to raise the issue of keyword search terms. The general counsel tells you that search terms are work product and to tell opposing counsel to issue document production requests and an appropriate response will be made. She then leaves to rush to a meeting.

At the Rule 26 meet-and-confer session, opposing counsel surprises you. She proposes to produce documents resulting from a search for electronic documents based on predictive coding, instead of using an approach based solely on a keyword search, and seeks your agreement. She refers to the results of TREC Legal Track and the decision of Da Silva Moore v. Publicis Group & MSL Group, 2012 U.S. Dist. LEXIS 23350 (S.D.N.Y. Feb. 24, 2012) to legitimize her client’s proposed search approach.

You have never heard of TREC Legal Track but don’t want to admit it. You have never heard of the Da Silva Moore case either but despite your ignorance, you stay cool. You dig in your heels and refuse to agree. Opposing counsel threatens to seek from the court approval to recover the costs of human review if keywords solely are used to identify potentially responsive documents. You stand your ground. You say nothing to your client about this part of the Rule 26 conference.

At the Rule 16 conference, the district court hears from both sides and scolds you for refusing to cooperate with your opponent. The district court gives your opponent the option to search for and produce electronically stored information based solely on predictive coding. Fearing more wrath from the judge, you say nothing further.

As discovery proceeds, you bone up on computer-assisted review and learn a lot more about how to use computer search tools to find relevant documents in a lawsuit. You start badgering opposing counsel with demands for more information about the sample set of documents being used to code the issues in the case and how opposing counsel is making relevance determinations. In the meantime, to fend you off, opposing counsel is listening to the advice of her e-discovery vendor and parroting whatever she is being told. A battle is looming over the size of the sample, the representativeness of the sample documents, coding decisions made by opposing counsel or her vendor (it is not clear who is calling the shots), and your demands for transparency.

What ethical issues might accompany these facts?

Before answering this question, this paper explores the jurisprudence of “cooperation” and “search technology” including the perilous waters lawyers will encounter if they do not comprehend the pitfalls that can accompany e-discovery stipulations. Then it will review the results of the National Institute of Standards and Technology’s Text Retrieval Conference “Legal Track,” which for the past four years has invited participants to test search technologies on mock production requests from known document sets in the hopes of finding ways to improve information retrieval at a lower cost than might be provided by manual review, yet as effectively or more effectively. That discussion will be followed by a review of two matters. One is Da Silva Moore, a case that explains what predictive coding is and that has created e-discovery news headlines for being the first reported decision to have a meaningful discussion of this form of computer-assisted review. The other is Kleen Products LLC v. Packaging Corp. of America, Inc., Civ. No. 1-10-cv-05711 (N.D. Ill.) where the court followed a very different path highlighting the difference between cooperation and capitulation. As will be explained, the court changed the dialogue from plaintiffs’ demand that the court order defendants to engage in predictive coding instead of a keyword-based search, to a defendant’s right to select the means it wishes to use to make a production. The paper then discusses upcoming “technology” amendments to the Model Rules of Professional Conduct. With this extensive backdrop, the stage will have been set for an analysis of ethical issues associated with this hypothetical.

More on the Ethics of E-Discovery: Predictive Coding and Other Forms of Computer-Assisted Review

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply