1. Federal and State Courts Issued Decisions Approving the Use of “Computer-Assisted Review” AKA “Predictive Coding” or “Technology-Assisted Review.”
Last year, numerous e-discovery commentators and vendors published articles and on-line resources and held seminars explaining how “computer-assisted review” (also known as “predictive coding” or “technology-assisted review”) works and discussing the expected benefits of using computer-assisted review. EDRM, which previously published the Electronic Discovery Reference Model depicting the various stages of the e-discovery process, published a draft reference model describing the process of computer-assisted review. The draft model defined computer-assisted review as a “process of having computer software electronically classify documents based on input from expert reviewers, in an effort to expedite the organization and prioritization of the document collection.” The major steps in the process according to EDRM are: setting outcome goals; setting the protocol for training the software; educating the human reviewers; coding documents by human reviewers; predicting results by the software applying the coding decisions; testing results through statistical sampling; evaluating results and repeating the process if necessary; and achieving the goals of the review. Additionally, The Grossman-Cormack Glossary of Technology Assisted Review was published. The glossary attempts to provide “a common framework and set of definitions for use by the bar, the bench, and service providers.”
A study from the RAND Corporation grouped the costs associated with e-discovery into three main categories: collection (8% of costs); processing (19% of costs); and review (73% of costs). Based on its study, RAND believed that companies could lower the cost of large document review projects by using computer-assisted review to reduce the number of documents requiring human review.
With all the discussion surrounding computer-assisted review, U.S. Magistrate Judge Andrew J. Peck became the first judge to issue an opinion that approved the use of that process to find and produce relevant electronically stored information (ESI) during discovery. In Monique Da Silva Moore v. Publicis Groupe & MSL Group, ___ F.R.D. ___, 2012 WL 607412 (S.D.N.Y. Feb. 24, 2012), Magistrate Judge Peck recognized that “computer-assisted review is not perfect” but encouraged counsel to consider using this technology to search for relevant ESI in appropriate cases: “What the Bar should take away from this Opinion is that computer-assisted review is an available tool and should be seriously considered for use in large-data-volume cases where it may save the producing party (or both parties) significant amounts of legal fees in document review.” Id. at *12.
Not long after Magistrate Judge Peck’s decision, a Virginia state court allowed the defendants “to proceed with the use of predictive coding for purposes of the processing and production” of ESI. In Global Aerospace Inc. v. Landow Aviation, L.P., No. CL 61040 (Va. Cir. Ct. Apr. 23, 2012), the defendants estimated they had 250 gigabytes of reviewable ESI from their computer systems, which could easily equate to more than two million documents. According to the defendants, a linear manual document review would take 20,000 man hours, cost $2 million, and locate only 60% of the potentially relevant documents. Keyword searching could be more cost-effective, but it may retrieve only 20% of the potentially relevant documents. Defendants believed that predictive coding, however, could be done in less time for less money and could retrieve upwards of 75% of the potentially relevant documents. (It has been reported recently that the defendants have completed the predictive coding process and that the plaintiff has not objected to the results.)
In a twist on what happened in the Global Aerospace case, it was the plaintiffs that sought to compel the defendants to use a form of predictive coding instead of keyword searching in an antitrust lawsuit pending in the Northern District of Illinois. In Kleen Products LLC v. Packaging Corp. of America, No. 10-5711, 2012 WL 4498465 (N.D. Ill. Sep. 28, 2012), after the defendants had already begun producing documents, the plaintiffs argued that the defendants’ search methodology was likely to find less than 25% of responsive documents but that the plaintiffs’ proposed “content-based advance analytics search” would find more than 70% of responsive documents at no greater cost. According to the plaintiffs, their proposed search methodology would provide more accurate results than Boolean keyword searches because it would identify relevant concepts out of the documents instead of focusing on matching words. U.S. Magistrate Judge Nan R. Nolan conducted two full days of evidentiary hearings regarding the parties’ positions. Magistrate Judge Nolan observed at the end of the second day that “[r]esponding parties are best situated to evaluate the processes, methodologies, and techniques appropriate for preserving and producing their own” ESI. 2012 WL 4498465, at *5 (citing The Sedona Conference, The Sedona Conference Best Practices Commentary on the Use of Search and Information Retrieval Methods in E-Discovery, 8 Sedona Conf. J. 189, 193 (Fall 2007)), and she urged the parties to find a middle ground that would provide reasonable assurance to the plaintiffs that they were receiving a high percentage of responsive documents without completely scrapping the defendants’ search methodology. Kleen Products, 2012 WL 4498465, at *5. After months of subsequent status conferences, the plaintiffs withdrew their demand and agreed that they would not argue that the defendants should use the plaintiffs’ proposed methodology for any document requests served before October 1, 2013. The parties agreed to meet and confer regarding an appropriate search methodology to respond to future document requests.
In In re Actos (Pioglitazone) Products Liability, MDL No. 11-2299 (W.D. La.), the parties in multi-district litigation pending in the Western District of Louisiana agreed to a protocol for a “Search Methodology Proof of Concept” to test predictive coding as a possible method for the search and production of relevant ESI collected from one of the parties. The protocol set forth a series of detailed steps and quality controls designed to train the predictive coding software on a smaller subset of the party’s ESI with the goal of eliminating the need to manually review documents below an agreed-upon relevancy score. Each side agreed to have three individuals work collaboratively as “experts” to train the software. The protocol built in numerous meet-and-confer checkpoints throughout the process, including at the end of the process where the parties were to meet and confer about finalizing the method for searching documents on a going forward basis.
Finally, the Delaware Court of Chancery took a different approach and sua sponte ordered the parties to use predictive coding even though neither party asked to use it. The court stated: “This seems to me to be an ideal non-expedited case in which the parties would benefit from using predictive coding. I would like you all, if you do not want to use predictive coding, to show cause why this is not a case where predictive coding is the way to go.” EORHB, Inc. v. HOA Holdings, LLC, No. 7409 (Del. Ch. Oct. 15, 2012). “[I]nstead of burning lots of hours with people reviewing, it seems to me this is the type of non-expedited case where we could all benefit from some new technology use.” The court also ordered the parties to agree to use the same e-discovery vendor to house their documents, stating, “If you cannot agree on a suitable discovery vendor, you can submit names to me and I will pick one for you.”
2. The American Bar Association Passed New Model Ethics Rules Regarding Technology.
While many lawyers and judges were learning about computer-assisted review last year, the American Bar Association announced changes to its Model Rules of Professional Conduct to address the impact of technology on the practice of law. A new comment to Model Rule 1.1, which defines a lawyer’s duty of competence, now makes clear that lawyers have an obligation to have a basic understanding of relevant technology: “To maintain the requisite knowledge and skill, a lawyer should keep abreast of changes in the law and its practice, including the benefits and risks associated with relevant technology, engage in continuing study and education and comply with all continuing legal education requirements to which the lawyer is subject.” Another new comment to Model Rule 1.1 lists the factors that lawyers should consider before retaining or contracting with other lawyers outside their own firm, including: the education, experience, and reputation of the nonfirm lawyers; the nature of the services to be performed; and the legal protections, professional conduct rules, and ethical environments of the jurisdictions in which the services will be performed, particularly relating to confidential information. Together, these changes will likely impact how lawyers plan for and manage large document review projects. Lawyers will need to stay current with available search and review technologies (e.g., predictive coding, near duplication, email threading, etc.), and closely analyze alternative staffing models such as using contract attorney located outside the United States.
Model Rule 4.4, which addresses the rights of third persons, now expressly includes ESI within the scope of a lawyer’s obligation to notify the sender of documents mistakenly sent or produced by opposing parties or their lawyers. The commentary states that the receipt of metadata in electronic documents creates an obligation only if the receiving lawyer knows or reasonably should know that the metadata was inadvertently sent.
Model Rule 1.6(c), which addresses client confidentiality, now...
