Lawyers have become somewhat comfortable with the notion that they can use predictive coding in eDiscovery when it best serves the interests of their clients. Such a position is not really controversial since it is consistent with the overarching eDiscovery doctrine that a responding party is best situated to determine how it should search for, review, and produce its responsive documents. Some courts are bucking this trend, however, and unilaterally ordering parties to use predictive coding. They are apparently doing so to address limitations with keyword searches and parties’ inability to cooperatively resolve disputes over keywords.
For example, the court in Aurora Cooperative Elevator Company v. Aventine Renewable Energy ordered the parties to use predictive coding in connection with the first phase of discovery in that case. In doing so, the court resolved the parties’ feud regarding the nature and extent of keyword searches to locate responsive emails from certain of the defendants’ custodians. While the court generally allowed the parties to negotiate the nature and scope of the predictive coding process, it also compelled the parties’ protocol to satisfy certain requirements, including the designation of an expert to design a computerized search to examine the parties’ electronic records.
Similarly, the court in Independent Living Center of Southern California v. City of Los Angeles ordered the defendant city to use predictive coding to produce “the 10,000 most relevant documents” from certain of its databases within “six weeks.” The court turned to predictive coding to eliminate continued haggling between the parties over keyword searches and to expedite the discovery process. In connection with that order, the court gave the parties latitude to execute a predictive coding use protocol. However, the parties were unable to do so – “[a]s with many of the other issues in this case, however, counsel for Plaintiffs and counsel for the City have not agreed on the protocol to be used” – which left the court to fill in many of the details. The court did so by ordering the parties to jointly develop “the relevance screening process and the training process for the predictive coding system,” prohibiting the city from using documents it previously identified by keyword searches to train the predictive coding workflow, and making the plaintiffs financially responsible for any documents that they might seek to discover beyond the production limit of 10,000 documents.
For many attorneys, the Aurora Cooperative and Independent Living Center cases might be disconcerting. Since the “majority of lawyers in trial practice today have never laid eyes on a predictive coding tool,” being forced from the familiar comfort zone of keyword searches into the new realm of predictive coding may seem like “mission impossible.” Moreover, many counsel would understandably recoil from court orders that remove their discretion to carry out the document production process. Even experienced predictive coding practitioners might balk at the nature of the terms the courts imposed regarding the use of predictive coding.
This backdrop of complexity underscores the need for lawyers to become familiar with predictive coding and the guidelines for using it in an eDiscovery workflow. This step is consistent with the direction from the ABA and other bar organizations, which now require lawyers to understand “the benefits and risks associated with relevant technology” such as predictive coding so they can appropriately advise their clients. Such a task should not be seen as daunting. Various organizations such as the Coalition of Technology Resources for Lawyers have promulgated materials that provide clarity and direction on these issues, including its recently released Guidelines-Regarding-the-Use-of-Predictive-Coding.
Finally, lawyers – regardless of their experience level with predictive coding – would be well advised to read the above referenced decisions and avoid where possible the scenarios that led to the orders in question. Understanding the playing field from the court’s perspective will undoubtedly help counsel better frame the issues and could ultimately generate results more to the court’s liking and to the benefit of clients.
*Special thanks to Recommind for allowing the CTRL Initiative to repost this fantastic guest blog!