LegalTech New York Graduates from “ECA” to “Assisted Technology Review"
If you thought descending on New York was going to be a vacation, you are probably feeling otherwise by now. I am confident that the recuperation process has begun after another successful LegalTech Show. If you have ever been before, this year’s show had the typical “glitz and glamour” as past shows. Last year, the talk of the town was Early Case Assessment (“ECA”) and nearly every workshop or presentation incorporated the topic in some shape or form. In contrast, this year’s show nearly made no mention of ECA. Instead, the unofficial theme of the Show was “Assisted Technology Review,” otherwise known as Predictive Coding, or more recently referred to as Predictive Priority TM.
The first plenary session titled, “Man vs. Machine: The Promise/Challenge of Predictive Coding & Other Disruptive Technologies” featured Maura R. Grossman (Litigation Counsel at Wachtell, Lipton, Rosen & Katz), Ralph Losey (Partner and National eDiscovery Counsel for Jackson Lewis), and The Honorable Judge Andrew Peck (U.S. Magistrate Judge for the Southern District of New York). The title of this session was somewhat misleading because by the end of the presentation, the conclusion was clearly a resounding “machine” as the victor, with little to no discussion about the “man” side of the equation.
Future blog posts will dissect the contents of this panel more in-depth because there was a lot of content discussed, yet one major takeaway is a baseline definition of “Technology Assisted Review” (“TAR”) which was defined as:
…an umbrella concept that involves keyword search, conceptual search, clustering, relevance ranking, sampling and predictive (aka computer-assisted) coding…[meaning] tools that use sophisticated algorithms to enable the computer to determine relevance, based on interaction with (i.e. training by) a human reviewer. - Man vs. Machine Plenary Session
This definition takes into account the need for the “human reviewer” to be a seasoned partner or team that codes a “seed set” of documents that the computer will use to prioritize the review and/or determine responsive versus non-responsive documents. Grossman described the process with an analogy of an ophthalmologist’s examination, where an individual’s eyes are tested with lenses on the left and right side to determine the appropriate strength of the prescription. Eventually, it will become challenging to differentiate between the lenses and that is when your prescription has been identified. Similarly, the “high level trainers” as Losey referred to them, will train the computer to identify the appropriate documents to review and produce.
While the bulk of the show focused on TAR, very little emphasis was placed the importance of where TAR would be most effective and where it should appear in the Electronic Discovery Reference Model (EDRM). I argue that it should occur where LegalTech left off last year, in ECA. ECA allows you to access the entire corpus of documents and infuse your seed set at the beginning, before it gets to the review phase. Then you can feel more comfortable with the integrity of the data set that will be used for the traditional manual review. Stay tuned, for there is more to come from LegalTechNew York!