In a new development on the predictive coding front, a federal court just issued an order under Federal Rule of Evidence 502(d) that “encouraged” the parties “to conduct computer-assisted privilege and work product review of ESI prior to production in order to expedite review and speed production.” Although there are some 30 opinions to date discussing and endorsing the use of computer-assisted review, this case appears to be among the first to specifically focus on the use of such technology in the context of privilege reviews.
The case at issue—Good v. American Water Works—is part of the class action litigation surrounding the now infamous Freedom Industries chemical spill that tainted the drinking water for around 300,000 West Virginians. The suit, instituted in January, has progressed quickly into discovery and the parties had agreed on all but one issue in the discovery protocol regarding privilege review and 502(d) clawbacks. Naturally, the plaintiffs wanted a faster review process and were willing to enter into a broad 502(d) arrangement in exchange for the exclusive use of technology-assisted review to expedite production. However, if the defendants intended to proceed with a manual privilege review, the plaintiffs argued that only 502(b) protection should apply.
The Difference a Letter Makes The reason for the dispute is in the differences accorded to the production of privileged materials under Rule 502(b) and (d). In short, 502(b) is the general “inadvertent waiver” rule that is in effect in all cases and protects against waiver under a multi-factor analysis. The problem with 502(b) is that its protection hinges on an inherently unpredictable analysis of whether the privilege holder “took reasonable steps to prevent disclosure.” In contrast, 502(d) orders provide far broader protection against privilege waivers. Such orders generally protect privilege holders regardless of whether a disclosure was inadvertent or the preventative steps a holder took were reasonable.
A Nuanced Argument The plaintiffs tried to frame the issue as a dichotomy. If the defendants wanted to proceed with manual review for privilege, then they would naturally be taking the reasonable steps to prevent disclosure required for 502(b) protection and would not need a 502(d) agreement. On the other hand, if the defendants were willing to produce documents without any manual review, the plaintiffs would reward them with a 502(d) agreement. The plaintiffs contended against coupling a 502(d) order with manual review:
Plaintiffs can see no practical reason for Defendants to engage in any kind of manual privilege review prior to production of documents other than to delay the production of documents. In fact, Plaintiffs can see no very good reason for any kind of privilege review at all prior to production, but are willing to agree to computer-assisted searches and other machine-based privilege reviews as a compromise because those are unlikely to result in any significant delays.
Ultimately, the court refused to go that far and order the exclusive use of computer-assisted review, but it did preserve that option in case the defendants miss their production deadline. Further, in the published 502(d) order, the court encouraged defendants to use computer-assisted review in combination with linear review.
The So-What Factor? This isn’t just just another case on the growing list of courts to approve and endorse the use of predictive coding. The fact that the court left the exclusive computer-assisted review option on the table is substantial. The likelihood of the defendants missing a deadline in this litigation, which is just one of several related to the class action and bankruptcy, is certainly not de minimis. We may be revisiting this case in a month or two.