Predictive coding technology is a lot like the popular poker game Texas Hold ‘em. Both can be risky and expensive for players who don’t understand the fundamentals of the game. Good players understand what kind of information they need from their opponents in order to make informed decisions. Bad players, on the other hand, ignore important elements of the game like statistics that must be understood in order to avoid making big mistakes.
In January, Maura R. Grossman, Craig Ball, Ralph C. Losey and I (Matthew Nelson) discussed these and other parallels between poker and predictive coding in front of a full-house at Legal Tech New York (LTNY). Please enjoy some of the live video clips from part one of our session titled, “How Good is Your Predictive Coding Poker Face?” as you contemplate whether or not you’re ready to go all-in with predictive coding technology.
Why “reading” your opponent is important
Recognizing your opponent’s strengths and weaknesses, aka “reading your opponent” is a key strategic consideration whether you’re playing poker or establishing a predictive coding protocol. In litigation, the Federal Rule 26(f) discovery conference often serves as the best opportunity to evaluate your opponent’s eDiscovery acumen. What if your opponent isn’t tech savvy? Do you still have a legal or ethical obligation to explain what kind of technology you plan to use during discovery? If not, should you disclose what technology is being used anyway? Watch this video clip as the panel examines whether or not opposing counsel’s level of technological sophistication is a factor that should be considered when deciding whether or not to reveal your technology approach.
Is the game changing?
Although the rules of poker are constant, the way the game is played continues to change and evolve in sophistication as more and more players try different strategies and approaches. Similarly, many believe a new eDiscovery paradigm is developing whereby methodologies for responding to discovery are likely to be more closely scrutinized by the court and opposing parties than in the past? Craig Ball thinks that responding parties have been “getting away with murder” for a long time and that the eDiscovery game is changing. Ralph Losey believes in Sedona Principle 6 and the notion that responding parties are in the best position to understand their data regardless of whether or not the game is changing. Take a look at how the panel plays this tricky hand.
Should a request to see the predictive coding “discard pile” be treated as a stone cold bluff?
Do parties have an obligation to disclose non-responsive files (aka the discard pile) used to train the predictive coding system? What if the opposing party insists on requiring the disclosure of the discard pile as part of the predictive coding protocol? Is your opponent bluffing or should you think seriously about cooperating with the request? If you cooperate, will too much transparency lead to a gradual erosion of traditional work product protection?
How Good is Your Predictive Coding Poker Face? (Part Two)
Stay tuned for part two of “How Good is Your Predictive Coding Poker Face?” where our panel digs deeper into the parallels between poker and predictive coding and considers the possibility of a “middle ground” approach that may satisfy both parties. The panel also explores other interesting issues like the importance of statistics, the need for transparency, and the dwindling role of keyword search technology in eDiscovery.
In the meantime, let us know what you think. Is predictive coding changing the eDiscovery game? Are producing parties getting away with murder? Should attorneys be required to show their predictive coding discard pile to the other side if they use predictive coding?
