Predictive Coding: a primer for electronic discovery
November 15, 2010| Law & Forensics
Part 1 of 1
Daniel B. Garrie, is the Senior Managing Partner at Law & Forensics LLC and works out of the Seattle, Los Angeles, and New York offices. He focuses on e-discovery, digital forensics, cyber security and warfare, data privacy, and predicitive coding working with law firms, governments, companies, and non-profits globally.
The influx of unstructured data only compounds the complexity of electronic discovery. The use of keyword searches to identify relevant data was appropriate in the past, but there is simply too much data and too little time for the lay In-house and outside counsel should evaluate the benefits of applying predictive coding to ongoing litigation.
Predictive coding is software that ranks documents according to relevancy. Through algorithms, an attorney can “teach” the software the key words, concepts, and phrases that he or she is looking for. After “learning” what the attorney is searching, the software separates or “codes” the documents according to the attorney’s terms. As a result, counsel can save its client time and money without sacrificing accuracy.
Predictive coding deserves to become the new default in discovery, as absolute perfection is neither necessary nor required. For example, in [Pension v. Banc of Am. Sec. LLC], 685 F. Supp. 2d 456, 461 (S.D.N.Y. 2010), Judge Shira Schenidlin stated in dicta, “courts cannot and do not expect that any party can meet a standard of perfection” with respect to the production of ESI. “Nonetheless,” Judge Schenidlin continued, “the courts have a right to expect that litigants and counsel will take the necessary steps to ensure that relevant records are preserved when litigation is reasonably anticipated, and that such records are collected, reviewed and produced to the opposing party.” Attorneys can reasonably infer that predictive coding is more than likely defensible as long as counsel makes reasonable steps to review and produce responsive documents to the opposing party.
The Federal Rules provided guidance of what is likely reasonable production results for opposing counsel. Federal Rule of Civil Procedure 26(g)(1)(A) states that “by signing [disclosures], an attorney or party certifies that to the best of the person’s knowledge, information, and belief formed after a reasonable inquiry: (A) with respect to a disclosure, it is complete and correct as of the time it is made[.]” What then is a “reasonable inquiry”? It will depend on the circumstances, and those circumstances may be defined by the amount of ESI in relation to the amount of responsive documents produced.
Furthermore, predictive coding may be just as good as or even better than an attorney conducting document review. Maura Grossman and Gordon Cormack have attempted to prove this. , [“Technology-Assisted Review in E-Discovery Can Be More Effective and More Efficient Than Exhaustive Manual Review,”] http://jolt.richmond.edu/v17i3/article11.pdf. They provide evidence that a “technology-assisted process,” such as predictive coding, “can yield higher precision in the document review process than the way in which humans examine documents.”
If opposing counsel objects initially, both parties can select a vendor and agree in writing to use the technology. The producing party has the burden of creating a cooperative spirit and may want to request a demonstration sample to prove the accuracy and reliability of predictive coding. If opposing counsel still opposes the use of predictive coding, the producing party should request that opposing counsel pay for the cost of manual review.