How AI is Reshaping Privilege Review

How AI is Reshaping Privilege Review

The argument for applying AI technology to automate routine, repetitive legal work is borne by cost efficiencies. But what about high-stakes litigation activity like privilege review where a matter can turn on a single document that has been overlooked in the review process?

Text IQ assembled a who’s who of eDiscovery thought leaders for the third in The Inevitable 2020 Series. The topic of conversation turned on the question: can AI prove to be more effective for first level review of privileged documents by understanding communications in context, and help legal teams be more efficient in focusing on documents that require their judgement calls? And, if so, what are the implications for the litigation process?

 

Watch Now

Our panelists:

● The Honorable Andrew J. Peck, who served for 23 years as a United States Magistrate Judge for the Southern District of New York including a term as Chief Magistrate Judge, is Senior Counsel at DLA Piper. Judge Peck has worked as special discovery counsel and serves as an arbitrator, mediator, and Special Master.

● IQVIA General Counsel, Laura Kibbe is a 2006 Corporate Counsel Magazine’s “Top 10 Innovative In-House Counsel,” as well as “Trailblazer” in electronic data. Laura holds a US patent “for the system and method of reviewing and producing documents utilizing technology assisted review.”

Bobby Malhotra is eDiscovery counsel with Munger, Tolles & Olson. A member of the firm’s litigation practice group, Mr. Malhotra develops innovative, cost-effective, and defensible strategies for the preservation, collection, review, and production of electronically stored information.

● Recognized by Super Lawyers for his work in e‑discovery, Scott Reents is Cravath, Swaine & Moore LLP’s Data Analytics and E‑Discovery Lead Attorney. Scott is expert in the use of advanced technologies such as technology assisted review (TAR).

Traditional TAR vs. AI Designed for Privilege Review

“The challenge of using TAR/AI in privilege review is... that you need to minimize the risk of error. And so when we use Text IQ or other technology approaches to privilege identification, we're not looking for 80% recall, we're looking for 99.5% plus recall. The risk tolerance for missing privilege is practically zero.” -- Reents

 

Munger, Tolles’ Malhotra agrees. He notes that while available off the shelf Technology Assisted Review (“TAR”) options work for relevancy review, they don't do a great job dealing with parsing gray areas such as legal versus business advice or understanding who's communicating with whom. Technologies like Text IQ’s AI for Legal focus on relationships, which is especially important.

“Since my DaSilva Moore case in 2012 where I, for the first time ever in the world said that computer assisted review...what's now TAR, will be acceptable in the right case...[And,] relying on Sedona Principle 6, that the responding party is in the best position to determine how it is going to review material for relevance production...that pretty much is the state of the law on using TAR for relevance and responsiveness...

There hasn't been a case where AI for privilege review has been challenged and ruled upon by the court... or anything else like it in privilege review.” -- Judge Peck

Judge Peck talking about defensibility of AI in privilege review

A Stark Choice

Cravath’s Reents candidly opines that the choice is not between spending three months and dozens of attorneys carefully reviewing every single document for privilege versus using some cheaper AI alternative. The choice is rather, “we have four weeks to do this. How are we possibly going to do this in four weeks?”

Laura Kibbe adds wryly: “A gun to your head always helps.”

“This may be particularly true in a Hart-Scott-Rodino (HSR) second request” responds Judge Peck where “you don't have even the leeway you might in a litigation of the same size to take a longer period of time to do your production.” The value of AI in HSR Second Requests is a topic of conversation covered in depth in The Inevitable 2020 webinar When the Odds Are Against You: Antitrust Merger Review.

The Desired Future State

Could AI-driven review and classification further help legal teams as the technology matures and is adapted to new use cases? The potential lies in reducing the time taken to deal with matters through automation, and enabling lawyers to focus on questions of law rather than routine review.

“My utopia is to take all of my documents...throw them in the hopper and I want different buckets to come out. I want a responsive...but not what’s privileged bucket to come out. I want ‘here's all your other special treatments, i.e. PII, PHI kind of thing so that you're tackling the collection as a collection, and you're learning how everything relates.'” -- Kibbe

“[It’s] allowing lawyers to do what lawyers do best which is to make those judgment calls through the legal research and not what they're frankly over-priced to do, which is just sort of mechanically do the same thing over and over and over again. That's, that's what technology is good at. And so there's a natural symbiosis there.” -- Reents

“That's such an important point. That's what I really see happening in the future kind of large-scale adoption of these AI models...I really do think that is the wave of the future.” -- Malhotra

Check out the full discussion here and peruse the full Inevitable 2020 series line up here