March 11, 2015





SDNY Recognizes Party Using TAR Not Required to Disclose Control Sets to Opposing Party
By Wendy Butler Curtis and Dan Guerra

Magistrate Judge Andrew J. Peck’s recent decision in Rio Tinto PLC v. Vale S.A., S.D.N.Y., No. 14 Civ. 3042 (RMB)(AJP) (March 2, 2015), recognizes disclosure of non-responsive documents is not a per se requirement in cases where TAR is used. This dicta in the opinion reflects an important evolution in Judge Peck’s jurisprudence on this issue and generally in the case law.   

Three years ago, in Da Silva Moore v. Publicis Groupe & MSL Grp., 287 F.R.D. 182 (S.D.N.Y. 2012) (Peck, M.J.), aff’d, 2012 WL 1446534 (S.D.N.Y. Apr. 26, 2012), Judge Peck authored the first decision approving the adoption of a detailed TAR protocol. TAR, or Technology Assisted Review, is a means of combining human knowledge with technology to make discovery more efficient and cost effective. TAR begins with a subject matter expert coding documents (known as seed or training sets) to calibrate the computer algorithm. Once the algorithm is sufficiently trained, it is run across the entire universe of collected documents to categorize them as relevant or non-relevant. Various steps can then be taken to validate the accuracy of the results and to further process the documents to meet a party’s discovery obligations. Since Da Silva Moore, “the case law has developed to the point that it is now black letter law that where the producing party wants to utilize TAR for document review, courts will permit it.” Rio Tinto PLC, at *2. It is no longer a question of if TAR can be used to efficiently manage the burdens of large-scale eDiscovery; but rather, whether TAR protocols are necessary and if so, what degree of transparency is required.

As Judge Peck observes, “[o]ne TAR issue that remains open is how transparent and cooperative the parties need to be with respect to the seed or training set(s).” Id. at *4. After reviewing the case law on this point, Judge Peck concludes that the courts “are split and debate in the discovery literature is robust” when parties disagree about the level of transparency required for TAR’s use. Id. at *5.

In Da Silva Moore, Judge Peck noted a lack of unanimity on the topic while favoring enhanced cooperation and increased transparency between the parties:

While not all experienced ESI counsel believe it necessary to be as transparent as MSL was willing to be [by disclosing all non-privileged control set documents], such transparency allows opposing counsel (and the Court) to be more comfortable with computer-assisted review, reducing fears about the so-called “black box” of the technology.  This Court highly recommends that counsel in future cases be willing to at least discuss, if not agree to, such transparency in the computer-assisted review process.

Da Silva Moore, at 192. However, in Rio Tinto PLC, Judge Peck’s stance towards seed set transparency requirements appears to have evolved. While still professing to be a general advocate of cooperation, Judge Peck highlights the existence of other viable tools that can be used in lieu of transparency requirements:

[R]equesting parties can insure that training and review was done appropriately by other means, such as statistical estimation of recall at the conclusion of the review as well as by whether there are gaps in the production, and quality control review of samples from the documents categorized as non-responsive.

Rio Tinto PLC, at 6. This decision validates verification measures producing parties may offer as an alternative to opposing party requests for production of non-responsive seed set documents. When parties select a TAR service provider and review team, it is important to ensure they have experience with execution of and potential disclosure of these measures. In Rio Tinto PLC, Judge Peck does “not rule on the need for seed set transparency . . . because the parties agreed to a protocol that discloses all non-privileged documents in the control sets.” Id. Immediately after acknowledging this, Judge Peck succinctly summarizes the key point underpinning his apparent skepticism towards seed set transparency requirements. “One point must be stressed – it is inappropriate to hold TAR to a higher standard than keywords or manual review.” Id. The implications of those 20 words for TAR protocols in general and for seed set transparency disputes in particular cannot be ignored. In a traditional search term and human review, the producing party has no obligation to produce non-responsive documents. Holding a TAR review to the same standards precludes any argument that the producing party has an obligation to produce non-responsive documents; regardless of whether said documents are part of seed or training sets.

This case is an important development in TAR. Requesting parties have used the Da Silva Moore decision and comments by Judge Peck to suggest that TAR may not be used in a matter without allowing a requesting party input on tool selection, workflow and most importantly, review of non-responsive documents as part of the crafting of an agreed upon, complex and detailed TAR protocol. This decision, while still advocating for cooperation, moves the law closer to Sedona Principle 6 that states “[r]esponding parties are best situated to evaluate the procedures, methodologies, and technologies appropriate for preserving and producing their own electronically stored information.” The Sedona Conference, The Sedona Principles: Best Practices Recommendations & Principles for Addressing Electronic Document Production ii (Jonathan M. Redgrave et al. eds., 2nd ed. 2007). This permits a party’s decision whether or not to use TAR to be based on the “just, speedy, and inexpensive determination of every action and proceeding” precepts enshrined in the first rule of the Federal Rules of Civil Procedure, rather than a deliberation over the worth of entering into a potentially Faustian bargain.