Jessica Stephen, The Daily Record Newswire
Thanks to computers, the days of flipping through stacks of paper documents are over. But how much can attorneys rely on computers to decide what evidence is relevant and what is not? And what role can predictive search programs, like technology-assisted review, play in finding those few files among hundreds of thousands of documents?
Attempts to find answers to those questions surged three years ago, when U.S. Magistrate Judge Andrew J. Peck in the Southern District of New York issued the first ruling approving the use of technology-assisted reviews, otherwise known as TAR, in electronic discovery.
Interest swelled again earlier this year when Peck posited in another opinion that “it is now black letter law that where the producing party wants to utilize TAR for document review, courts will permit it.”
“People are fired up about it,” said Timothy D. Edwards, a national speaker on e-discovery and records management who practices in Madison, Wisconsin. “There are some cases that have suggested it’s the gold standard already, and others that are less enthusiastic. But how the courts are going to embrace it — whether a court will be critical for not using technology-assisted review — that’s the question.”
That may be true for the e-discovery experts who have been following TAR since it first came on the scene nearly five years ago. But for many attorneys who are just learning about TAR and its potential, the questions are more basic: What is it? How is it used? And why should we use it?
Those are questions Ralph C. Losey, national e-discovery counsel for Jackson Lewis, helps attorneys across the country answer.
Often, his explanation starts with mechanics: “[TAR] uses advanced analytics, sophisticated computer code that goes beyond keyword searches.”
When attorneys’ eyes start to glaze over, he tries this:
“Have you used Pandora, where you give a thumbs-up and down to a song and it teaches [the program] the kind of music you like? TAR is the same. You teach it: These are relevant, these are not.”
For the majority of attorneys, TAR is most useful for sifting through large amounts of digital discovery.
“The most common use is to try to sort your own clients’ documents into relevant or irrelevant,” Losey said. “It can also be used to sort and find privileged documents among the relevant documents, so you can log them and remove them before you produce them. The third way is to look at large volumes of documents that have been produced to you; if you have 100,000 or 200,000 documents, you have to find the ones that are useful to you. So, it’s used by both the producing and the receiving party.”
To many attorneys, it almost sounds too good to be true, Losey added.
“And I don’t really blame them,” he said. “But it really can be less expensive and more accurate if done correctly. With the computer helping me, I can review thousands of documents in an hour. Most people can review 50 at best. Plus, humans make mistakes. Computers, if they’re programmed right, will follow your instructions with total consistency. So, they’re not only fast, [but] they can read a million files in a minute and they’re totally consistent.”
Edwards agreed.
“When you’re dealing with large volumes of data, manual review — where a human being looks at each document to determine privilege issues or whether the document is relevant and needs to be produced — can be very time-consuming and very expensive. And there are inherent quality-control problems that can lead to a margin of error that is not acceptable,” Edwards said. “Of course, a computer is a little different. It will only do what you tell it to do. If it’s pointed in the right direction and you use appropriate search terms and protocols, they can be just as reliable, if not more reliable, than human review. … Neither system is perfect, but one is certainly less expensive if you’re dealing with large volumes of data.”
There are downsides.
“I think the main drawback is the search itself,” Edwards said. “Any time you use keywords or search terms, you’re confining yourself to a specific universe of documents. What that means then is the possibility that key documents will be missed that don’t include the search terms in question.”
When dealing with large amounts of data, Edwards advises working with opposing counsel to come up with a protocol.
“And work with someone who is skilled in the software platform,” he said. “Use it in a collaborative way with the assistance of someone who is proficient in the technology. It’s going to be 10 times more helpful than if one party does it on his own.”
Losey also urged attorneys to do some homework before buying a TAR program.
“You have to be very careful what software and what e-discovery you’re dealing with, and there can be a huge difference,” Losey warned. “It’s oftentimes, ‘You get what you pay for.’ … It takes time and effort to build really good software.”
And to learn how to use it.
“It is a bit laborious,” Losey admitted. “It’s like practicing law; it has a learning curve to it. You certainly can’t just push the predictive coding button. So, take your time. Pick your best and brightest, and have one attorney go deep with it. You can’t have every attorney become an expert in this.”
But it’s something lawyers should try, he added.
“It really can save your clients money. And it’s another part of practicing law, finding relevant evidence — just now most of the evidence is in a computer, not a filing cabinet,” Losey said. “You just have to train yourself to find these documents.”
—————
Jessica Stephen is a reporter for the Wisconsin Law Journal, a sister publication of Lawyers Weekly.
