On August 13th, Text IQ hosted its third webinar in The Inevitable 2020 series. This webinar, titled “How AI is Reshaping Privilege Review,” focused on the challenges in privilege review and how AI is overcoming them. The panel featured Judge Andrew Peck, a former U.S. Magistrate Judge for the Southern District of New York (1995-2018) and Senior Counsel at DLA Piper; Laura Kibbe, Assistant General Counsel at IQVIA; Bobby Malhotra, eDiscovery Counsel at Munger, Tolles & Olson LLP; and Scott Reents, Lead Attorney in Data Analytics and E-Discovery at Cravath, Swaine & Moore LLP.
- Privilege review is a time and resource intensive process that historically has been difficult to automate due to the difficulty of parsing gray areas. Recent advances in AI have solved these problems, and we are now able to automate 50-80% of privilege review, which was historically one of the biggest costs of eDiscovery.
- “There has to be a better way. I would be lying if I said I had not experienced an eight pass review. That is the challenge, and even the best reviewer is going to get it wrong at least part of the time.” - Laura Kibbe, Assistant General Counsel, IQVIA
- “There is a distinction between the technological options that are available. There are tools out there that are more enhanced - have the tweaks to focus on relationships, context, and vernacular.” - Bobby Malhotra, eDiscovery Counsel, Munger, Tolles & Olson LLP
- AI-powered software is not designed to replace human judgement, but it does allow attorney resources to be focused on the most difficult decisions.
- "These are incredibly difficult distinctions, that two lawyers could argue about - and I’ve seen them argue about - for hours. What technology can do is sort of risk score things. So you can say these are the ones that are highest risk, most likely to be privileged, least likely to be privileged, and what that allows us to do as lawyers is to focus our lawyer time on the hard calls.” - Scott Reents, Lead Attorney in Data Analytics and E-Discovery, Cravath, Swaine & Moore LLP
- Although case law which speaks directly to AI that can automate privilege review is limited, several trends suggest that courts not only accept technology that performs this function, but may increasingly recommend or require it.
- Getting stakeholders on board may be challenging, but proof of concept tests (pilot programs) and budget modelling are convincing tools.
- “When you see the bottom line results and the difference it can make in the expense of litigation and in the accuracy of the information produced in a litigation, it’s just amazing. It’s a game changer.” - Laura Kibbe, Assistant General Counsel, IQVIA
- "Work up a budget on what it would cost to do a traditional privilege review. Work up another budget on what it would cost to use AI enhancements, and see the cost savings. I think dollars and cents matter a lot to stakeholders." - Bobby Malhotra, eDiscovery Counsel, Munger, Tolles & Olson LLP
- “There is no substitute for taking even a tiny sample of your own data, showing it to your stakeholders, and saying, ‘Wow! Look! It did find that document.’... I get them to the table with a budget, but I can’t close the deal if I can’t show them [the AI software is] accurate.” - Laura Kibbe, Assistant General Counsel, IQVIA
- Using an unsupervised machine learning tool like Priv IQ requires a shift in mindset.
- “It’s a paradigm shift from document review, which gets pushed down to the lowest level, young associates… I expect you as the partner to be the one doing this. To be involved in training the algorithm and the tool… Any use of technology is garbage in, garbage out. So we’ve got to change the way we get the inputs.” - Laura Kibbe, Assistant General Counsel, IQVIA
- In a big, contract attorney review, a lot of those decisions tend to get pushed down the road when it’s sort of too late and a lot of things have been produced. I like [AI] technology because it allows us and our smartest, best lawyers to be empowered earlier on in the process to make the hard calls.” - Scott Reents, Lead Attorney in Data Analytics and E-Discovery, Cravath, Swaine & Moore LLP