Robot Review

Will predictive coding win the trust of the courts?

In February, Above the Law declared predictive coding its "legal technology buzzword of 2011." More recently, as covered on EDD Update, the buzz became a deafening roar as one e-discovery provider, Recommind, announced a patent for its "predictive coding" technology. Understandably, other providers offering what they believed to be "predictive coding" tools were quite miffed and the notion of patenting predictive coding remains hotly debated.

Let's take a deeper dive into what "predictive coding" really means, explore some of its purported benefits, and delve into the biggest question for many practitioners and litigants: Is it reasonable and therefore defensible to use?

We'll start with what predictive coding is not. It is not the previously accepted paradigm of "linear" document review, which involves human reviewers manually reviewing and coding documents that might have been subject to some simple applications of limiting and organizing factors, such as custodian, date ranges and keyword terms. These reviewers are the legions of lawyers referenced by John Markoff in The New York Times' report, "Armies of Expensive Lawyers, Replaced by Cheaper Software," 03/04/11.

While linear review might offer value in matters with small amounts of electronically stored information (ESI) ­ — or other concerns that necessitate human eyes on every page — it can become unwieldy when facing huge volumes of data. Within the Electronic Discovery Reference Model (edrm.net) life cycle, document review is widely-believed to be the most costly component, because cases now can involve terabytes of data. If document reviewers analyze every document, page by page, with reviewers charging by the hour, that generates quite a bill.

To Continue Reading: Click Here
----------------------------------------------
Source: law.com
By: Farrah Pepper