Shortly after expressing his hope that a federal judge would encourage the use of predictive coding for e-discovery, U.S. Magistrate Judge Andrew J. Peck did so himself. In fact, he has ordered it in Da Silva Moore vs. Publicis Groupe et. al., a Second Circuit Title VII gender discrimination class action case. And, more specifically, he ordered the parties to use Recommind’s Axcelerate product.
Being the first of its kind, this mandate is sure to address some concerns about predictive coding. According to predictivecoding.com:
Predictive Coding is a court-endorsed process that combines people, technology and workflow to find key documents quickly, irrespective of keyword. Due to its massive accuracy and efficiency gains, Predictive Coding is revolutionizing how Early Case Assessment (ECA), analysis and document review are done. Predictive Coding has three components:
Case experts use Predictive Analytics to find key documents quickly and irrespective of keywords
Keyword-agnostic machine learning finds other relevant documents
Proven workflow with integrated sampling delivers results to a statistical certainty
Recommind developed Predictive Coding in partnership with some of the world’s leading enterprises and law firms. Recommind customers have been using Predictive Coding for the past 5 years.
Although it has the potential to substantially cut discovery costs and increase the accuracy of document review, it is a relatively new technology that many feel is not yet well vetted. As such, they will put off its adoption until the following questions are answered.
- Is it accurate?
- Is it easy to use?
- Will courts accept it or will it be hard to defend?
- Does it “protect” confidentiality and privilege?
- What is the cost?
- How does it work?
- What solution should I use?
Da Silva is certain to give us some answers. With discovery costs averaging over $1 million, it will be valuable to know what the cost is in this case. In addition, we might learn more about how it works, its accuracy in pulling relevant documents and leaving behind the irrelevant, and if it “protects” confidential and privileged information.
To the question of accuracy, predictive coding will probably never replace the human element completely. For example, traditionally it has not been able to identify whether a document is responsive, irrelevant, or privileged. However, with advances in the technology, this will be less of an issue.
With the clear advantages that automation brings, predictive coding is sure to become mainstream as prices fall, confidence increases, solutions become easier to use, and as courts accept it.