Several years ago, we were told that technology-assisted review (TAR) was the wave of the future and, by now, manual attorney review would likely be a thing of the past. Although TAR has gained some traction, it is not yet the standard for document review and it certainly hasn’t replaced attorney review as predicted.

In an August 2016 case, even one of the judiciary’s biggest proponents of TAR, Judge Andrew Peck of the U.S. District Court for the Southern District of New York, refused to order the use of TAR. In Hyles v. New York City, 2016 U.S. Dist. LEXIS 100390 (S.D.N.Y. Aug. 1, 2016), the plaintiff wanted the defendant to use TAR to identify documents relevant to the case, but the defendant wanted to use keywords. Despite stating that “TAR is cheaper, more efficient and superior to keyword searching,” Judge Peck remarked that it is not up to the court “to force cooperation or to force the responding party to use TAR.” While Judge Peck left open the possibility that “[t]here may come a time when TAR is so widely used that it might be unreasonable for a party to decline to use TAR,” it is clear that time has not yet come.

In an October 2016 case, In re Viagra Prods. Liab. Litig., the U.S. District Court for the Northern District of California followed Judge Peck’s reasoning in a similar discovery dispute. In this products liability case, the plaintiffs wanted the court to order Pfizer to use TAR to locate relevant information. See In re Viagra Prods. Liab. Litig., 2016 U.S. Dist. LEXIS 144925 (N.D. Cal. Oct. 14, 2016). Pfizer wanted to use search terms. The court could find no basis in the law to compel a party to use one method over another and thus denied the plaintiffs’ request. However, the court left the door open to revisit the issue if the plaintiff could later prove that Pfizer’s production was deficient.

While the In re Viagra court left open the possibility of revisiting the method of review after production, the U.S. Tax Court in Dynamo Holdings Limited Partnership v. Commissioner of Internal Revenue, 2016 WL 4204067 (U.S.T.C. July 13, 2016), in July 2016 made the parties stick to the review method they initially selected. In this case, the petitioners asked the court for permission to use TAR, which the court granted. The parties worked together to develop a collaborative predictive coding protocol. The commissioner even went so far as to code the seed and training sets. Post-discovery, the commissioner wanted the petitioners to produce all documents containing certain search terms that were not produced through the TAR process. The court declined to order a “search term based production,” since the plaintiff had agreed to the use of TAR and was very involved in the process. The court noted that discovery responses do not need to be perfect, just reasonable, and even human review can be inconsistent.

The lesson here is that if parties agree to a particular approach for discovery — be it search terms, TAR or some other method — they should not expect to later request an alternative approach unless there is a “meaningful shortcoming” in the production.