Judge David Waxse of the United District Court for the District of Kansas discusses technology-assisted review.
Judge David Waxse of the United District Court for the District of Kansas discusses technology-assisted review.

A group of experts weighed in with their thoughts on perhaps the hottest topic of all right now in the world of e-discovery, technology-assisted review (TAR), on the first day of InsideCounsel’s 12th annual SuperConference. 

With attorneys across the nation anxiously awaiting the final determinations in both the Da Silva Moore v. Publicis Groupe and Kleen Products v. Packaging Corp. of America cases, it only seemed fitting that a panel discuss the finer points of predictive coding. To this end, health care company Abbott Laboratories’ senior counsel and director of eiscovery and records management Jason Fliegel; Sidley Austin Partner Jeff Sharer; e-discovery solutions provider Daegis’ managing director, document review services Adam Farber; and Kansas District Court Magistrate Judge David Waxse shared their views on the matter.

Technology-assisted review is viewed as the next evolutionary step to solve the problems attorneys are facing in regard to discovery, and is based on human training of a computer to learn and predict how to tag unreviewed documents, followed by sampling those documents to gauge accuracy. The experts agreed that, when properly used, TAR potentially offers significant savings in both costs and time spent reviewing documents.

The first question people must ask when assessing whether or not TAR is appropriate for their project is what is the volume of documents that needs to be reviewed?

“One ‘disadvantage’ is that the machine needs time to learn,” said Fliegel. “You have to take the time and show it both responsive and nonresponsive documents. In cases where the document set is large enough to litigate, this is where you really can make the ROI (return on investment) argument to use TAR.”

Sharer noted that using TAR isn’t always a good idea when a case includes a large number of nontraditional data types, and attorneys need to rely on optical character recognition to turn those documents into a machine-readable format.

Judge Waxse agreed with Fliegel and Sharer, but pointed out that the case’s context also is of significant importance. “What lawyers tend to forget is that we try only a small percentage of cases,” he said, noting that somewhere between two-thirds and three-fourths of cases eventually settle, so TAR may not be worth the cost or effort.

Perhaps the most important talking point in the session was of cooperation between opposing parties. Judge Waxse opined that the best way to keep discovery costs down and expedite the process is to have cooperation between both sides.

“One thing you need to think about in huge document review cases is that everyone should create one database and pull together all relevant documents,” he said. “Then try to agree on one mutual vendor and give that vendor what you want searched for. This can only result in huge cost savings.”

For more on technology-assisted review and predictive coding, read the following articles posted by InsideCounsel.com’s Outside Experts:

4 lessons counsel can learn from Da Silva Moore

Getting defensive about predictive document sorting technology

Technology-assisted document review: Better than the alternatives

The good, the bad and the ugly of e-discovery keyword search

Predictive coding gets its day in court