As noted by Joe Palazollo in the Wall Street Journal Law Blog on June 17 (“Why hire a lawyer? Computers are cheaper”) and his follow up primer on predictive coding on June 18, “After Court Decisions, Clients Mull Swapping Lawyers for Machines,” judges are increasingly supportive of the use of predictive coding technology—also known as technology-assisted review or automated document classification—to help manage the cost of setting human (attorney) eyes on millions of pages of electronically stored information to find information relevant to a matter. Despite encouraging the use of this advanced e-discovery tool, the opinions stop short of taking technology-assisted review to the next level: supplanting other forms of review entirely.
As my colleague Amanda Jones explained in a recent article in Metropolitan Corporate Counsel, predictive coding can be a tremendously effective way to meet a client’s fundamental goal in e-discovery: identifying “as many responsive documents as possible while reviewing as few nonresponsive documents as possible, at a cost proportionate to the value of the case.” And where predictive coding can save clients the most money by supplanting human labor is in the preliminary stages of review: it can help sift through mountains of documents, bringing the documents most likely to be responsive to the fore. However, it is foolhardy to suggest that this technology can entirely replace other discovery tools and processes, including eyes-on review.
Instead of being a discovery panacea, technology-assisted review should be viewed as a helpful supplement to predecessor technologies and methodologies, such as keyword search. For example, keyword searching can help cull extraneous documents from a set for review and thus enrich the set of documents used to train predictive coding technology, as Magistrate Andrew J. Peck recently acknowledged in his recent Da Silva Moore v. Publicis Groupe opinion. In addition, keyword searches can help target specific concepts that might not turn up in random sampling, which can ensure a more comprehensive review.
Moreover, technology-assisted review is best used in tandem with lawyers, subject matter experts, and experienced e-discovery professionals: it can free them to perform more nuanced analysis and higher-order tasks that ensure sound case strategy. For example, lawyers are still required to perform the higher-level review of documents prioritized in the system—most often, these are documents most likely to be significant to the litigation or protected by the attorney-client privilege or work product doctrine. Additionally, linguists can help choose the right keywords to maximize the return of responsive results while minimizing the likelihood of overlooking important variants or other related terms. Furthermore, statisticians can play a role in validating the reliability and quality of search results, sampling throughout the process as well as demonstrating that the process is consistent, ensuring its defensibility. Finally, information technology specialists can help guide lawyers in the use of the right types of technology.
In short, predictive coding might be able to replicate—and in some cases, outperform—more basic first-pass document review functions that entail sifting vast quantities of documents into “in” versus “out” piles, but it certainly cannot replace the knowledge, frame of reference and expertise of seasoned lawyers and legal technology professionals who are required to help manage the discovery process to a successful result.
About the Author