eDiscovery: 'Bad hair day’ for predictive coding opponents

Setting the stage. In two earlier columns (March 16 and April 6), I discussed Da Silva Moore v. Publicis Groupe, 11 Civ. 1279 (S.D.N.Y. Apr. 26, 2012), concerning what has been referred to as the "ground breaking" entry of "predictive coding" (sometimes referred to as computer assisted searches adopting an electronically stored information, ESI, learning protocol to detect responsive information) on the eDiscovery stage. Da Silva Moore is the result of an innovative ruling by Magistrate Judge Andrew Peck. The effort was subsequently brought to a halt by the order of Peck's supervising Federal District Judge, Andrew Carter, granting plaintiffs' request that he consider second-guessing Peck. In order to convince Judge Carter to finally overrule Judge Peck, plaintiffs' counsel mounted a campaign against Peck (including seeking his recusal) that appeared to be personal in many respects and well outside what many would consider the ethical perimeters of legal opposition.

Carter's April 25 Decision. Carter was definite that, "Judge Peck concluded that under the circumstances of this particular case, the use of the predictive coding software as specified in the ESI protocol is more appropriate than keyword searching." In adopting Peck's orders and denying plaintiffs' objections, Carter said that he did not find "a basis to hold" that Peck's decision "is clearly erroneous or contrary to law." Carter was definite in making it clear that there "simply is no review tool that guarantees perfection." He noted that it had been acknowledged by Peck and the parties "that there are risks inherent in any method of reviewing electronic documents." Cautioning that, while perhaps "appropriate in certain situations," it is costly to conduct "manual review with keyword searches," Carter went further by pointing out that, notwithstanding a situation where all parties are willing to go along with manual review of documents, it "is prone to human error and marred with inconsistencies" created by the determination of counsel concerning the responsive nature of the documents in question (alleged to be some three million mail messages).

What the fight is really about. Blogger Christopher Danzig puts it this way: "It's new and scary" intimidating lawyers as well non-lawyers by "new, untested technology" where "it takes guts to be an early adopter." In this case, "predictive coding means that you are handing over discovery to a computer," which is "an intimidating concept," notwithstanding that "human document review is expensive, skull numbing, and not necessarily effective." Danzig quips that human document review makes it "certainly more satisfying to yell at a young attorney who screws up" than to take on a computer. He throws the final barb by comparing the potential "intense fights in nearly every case that seeks to employ" predictive coding similar to those discovery contests already occurring demonstrating that it "is just life as usual."

The Da Silva Moore decision now has company. The floodgates are open. Kleen Products, LLC, et al. v. Packaging Corporation of America, et al. (N.D.Ill.E.D., April 8, 2011) has followed Da Silva Moore but there is now another twist. Global Aerospace Inc., et al. v. Landow Aviation, L.P. dba Dulles Jet Center, et al. (Va. State Court, April 23, 2012) just could be the first case where a court actually ordered the use of predictive coding. It has been reported that counsel for the defendant willingly signed the order but the plaintiff's counsel signed over the plaintiff's objection. It seems that trial lawyer and computer forensics and eDiscovery special master Craig Ball reportedly has cautioned that "the problem with an order like this is the devil is in the details,"e.g., the lack of discussion about quality control. Nevertheless, at least one eDiscovery lawyer has stated that he feels Global Aerospace has opened up the opportunity "for perhaps other litigants to be a bit braver about this issue."

Beware of the claims of service providers. Arkfeld warns that the claims of service providers (vendors) who advertise that their eDiscovery product is the best (especially in the new age of "predictive coding") require extra caution, and that we keep in mind the standards previously applied by courts concerning the use of computer-assisted review in the production process. In evaluating the appropriate nature of search terms and methods, "the court applies a reasonableness test to determine the adequacy of search methodology." In other words, could the search "have been expected to produce the information requested"? Here, Arkfeld cites Treppel v. Biovail Corp., 233 F.R.D. 363 (S.D.N.Y. 2006), but asks us to "always remember that, if the 'predictive coding' methodology does not capture the responsive information in a case and omits critical relevant information, then a party may subject themselves to sanctions." This is because, while the electronic searching of ESI can be valuable, "it must be used with a clear understanding of the strengths and weaknesses of the particular search and retrieval software and the qualifications of the person who is conducting the search."

So, much as is the case with life in general, there appears to be no full-proof remedy for all eDiscovery production ills. The fact that the courts are looking with approval at the new "predictive coding" efforts does not mean or guarantee that users of this approach will be insulated from the threat of possible sanctions simply because this method was employed by those who thought that it would be bulletproof.