News recently broke noting that Magistrate Judge Andrew Peck's recent opinion addressing predictive coding is in jeopardy of becoming obsolete less than a month after causing shockwaves within the eDiscovery community.
It seems the plaintiffs in the Da Silva Moore matter took exception to the procedural and temporal irregularities surrounding the seminal opinion; Judge Peck issued the opinion after plaintiffs filed their objections, thereby depriving them of the opportunity to object to the opinion itself. Consequently, plaintiffs objected and sought the opportunity to respond to the opinion itself to which Judge Andrew Carter obliged.
While this new twist may be significant in some respects, it does not mean Judge Peck's opinion is now meaningless, and it most certainly does not signal the end of the predictive coding movement and trend; if anything, this will only serve to draw more attention to predictive coding, the de jour subject in eDiscovery at the moment. Indeed, plaintiffs’ objections focused more on procedural and process issues, rather than the efficacy or validity of predictive coding in general or as a whole.
Judge Peck is clearly an advocate of predictive coding, and in issuing his seminal opinion endorsing, or at least agreeing to the use of, predictive coding, he was perhaps a little eager to make his point. However, that fact does not greatly diminish the power of what he said, and does not close the predictive coding door that Judge Peck opened (the opinion made a large splash for a reason). Despite this minor setback, the push for predictive coding will continue to move forward and the industry will continue to look to the Peck Da Silva Moore decision as a watershed moment regardless of the outcome of this new twist - you cannot unring the bell.
There were some peculiarities in Judge Peck's Order. Some of them had to do with his own lack of familiarity with using predictive coding. I was particularly surprised with the transcipt of the hearing where they attempted to decide how they would reach the populations the defense would produce. Unless I read it wrong, it didn't appear there was going to be any examination for false negatives. Which defeats the whole purpose of "training" the computer of what to look for.
ReplyDeleteStray_Bullet thank you for your comment. I agree with you that it was unusual that Judge Peck seemed willing to disregard, or modify, the "normal" training process for a more ad hoc approach. I get the feeling that he wanted to experiment with the system to see what would happen. Until predictive coding becomes commonplace and has proven its accuracy in numerous cases, I think there will continue to be some push back and fights over potential false negatives.
Delete