Many who consider Magistrate Judge Peck’s recent opinion and order in Rio Tinto PLC v. Vale S.A., which he titled “Predictive Coding a.k.a. Computer Assisted Review a.k.a. Technology Assisted Review (TAR) – Da Silva Moore Revisited,” will focus on his declaration “that it is now black letter law that where the producing party wants to utilize TAR for document review, courts will permit it.” We’ll revisit that statement in a moment, but first note that it is also black letter law that important discovery decisions get revisited. See, e.g., The Pension Committee of the University of Montreal Pension Plan v. Banc of America Securities, subtitled “Zubulake Revisited: Six Years Later.”
That said, Judge Peck’s opinion in Rio Tinto is important and is required reading, and his analysis justified revisiting Da Silva Moore, which we have previously discussed.
Transparency – Holding TAR to a Higher Standard
The opinion lays bare the key question practitioners face with machine learning tools: how transparent and cooperative must a party be when it chooses to use TAR? This important question remains unresolved. It is noteworthy that Judge Peck cites case law for the proposition that courts should not be in the business of dictating the review process a party “should use when responding to discovery.” And he observes, “It is inappropriate to hold TAR to a higher standard than keywords or manual review. Doing so discourages parties from using TAR for fear of spending more in motion practice than the savings from using TAR for review.”
Still, many hold (and will continue to hold) TAR to a higher standard because there is not widespread confidence that practitioners will deploy TAR correctly, or that a receiving party, unskilled in the use of machine learning tools (or even categorization and clustering tools), will be able to make or evaluate a “statistical estimation of recall at the conclusion of the review” or identify “gaps in the production.”
Even given the advances in case law and practice since Da Silva Moore, a large segment of bench and bar do not understand TAR, cannot explain it, and cannot defend the workflow they deployed in using it.
In a paper world, for better or for worse, no one would question the ability of a lawyer or paralegal to identify where the relevant paper lived and to manually review it and decide what is relevant. To be sure, even in a paper world some did better than others at finding what mattered. Yet we didn’t debate sharing the review criteria, identifying the people who applied that criteria, disclosing examples of documents deemed relevant and irrelevant, or the actual criteria for deciding when the identification and review of paper documents was complete.
But TAR tools measure imperfection. And while the discovery process has always been imperfect (and courts acknowledge it will never be perfect), bench and bar are still uncomfortable when faced with statistics that show a percentage of relevant documents that will not be reviewed or produced. TAR will be held to a higher standard than manual review until there are information retrieval process quality standards for counsel.
Stipulations – To File or Not to File
It is also noteworthy how Judge Peck handled the entry of the parties’ stipulation regarding the use of predictive coding in discovery. While we believe that a stipulation on the use of predictive coding is an important tool to reduce the likelihood of disputes about the outcome of the process, we remain reluctant to file the stipulation with the court (although we may reference the existence of a separate stipulation in court filings). Why?
- There is no requirement to file a predictive coding stipulation with the court.
- TAR technologies are developing and changing so quickly that a protocol a party is happy with today, for a particular case and TAR project, may not be the best protocol in a future case.
- The nature of the data and the search task facing me in one matter may suggest different approaches to transparency or different views as to the appropriate level of recall (a question not addressed by the parties’ stipulation in Rio Tinto).
Judge Peck appears to share our reluctance. He was appropriately cautious in noting that the stipulation he entered was negotiated by the parties, was not imposed by the court, and would not necessarily be appropriate in any future case.
Is it black letter law that “where the producing party wants to utilize TAR for document review, courts will permit it”? Maybe. Is it black letter law that the courts will view the outcome of every TAR process as complete and defensible? That question will be visited and “revisited” in opinions yet to be written.