KMWorld 2024 Is Nov. 18-21 in Washington, DC. Register now for $100 off!

Machine Learning Makes its Way (Slowly) into e-Discovery

Article Featured Image

Developing TAR guidelines

A working group of the Electronic Discovery Reference Model (EDRM) at Duke University, an organization that develops standards and guidelines for e-discovery, recently sought public comment on its draft guidelines for the use of technology-assisted review tools in discovery. “We need the community of lawyers and academics and in-house counsel and commercial entities to come together to define these things,” says James Waldron, director of EDRM at Duke Law and the former clerk of the U.S. Bankruptcy Court for the District of New Jersey. Making judges more familiar with TAR is key to the effort, he adds, because the magistrate judges have to call balls and strikes on discovery issues.

The EDRM guidelines are important, Gricks says, because lawyers don’t fully understand all the issues. “They need to have a decent resource that is going to let them recognize what is reality and what is not and the differences between tools and protocols so they can make decisions about using TAR and plan accordingly,” he explains.

Waldron notes that in addition to the guidelines, a TAR “best practices” work group is also readying a document for public comment.

The title of a March 2018 e-discovery conference at the University of Florida demonstrates the level of ambivalence in the legal community around TAR. Titled “Don’t Get Tarred and Feathered: Dealing with TAR, Keywords and AI,” the panel asked participants to debate whether “these tools and techniques could really replace keyword searching. Can they co-exist? Or are they just a waste of time and money?”

“I wouldn’t say the legal community is skeptical, but it is uninformed,” Gricks says. “The real issue that underlies the reluctance is that a good bit of the legal community is not informed on how the tools work and that leads to a natural reluctance to jump into TAR. You aren’t seeing people with experience saying that TAR tools don’t really work. It is more that people say they don’t know much about this and don’t feel comfortable with it.” 

KMWorld Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues