Construction lawyers understand
the importance of a comprehensive force majeure clause. But in an increasingly tumultuous world where
extraordinary events are difficult to predict, where can litigators find the
evidence to support (or refute) claims of impossibility or impracticability? Increasingly, the answer is in their client’s
own data. And with data volumes
increasing on a daily basis lawyers are turning more frequently to predictive
coding as their tool of choice.
Predictive coding, at the most basic
level, is a machine learning system where exemplar documents are provided and
the system retrieves other documents that are conceptually similar to the
exemplar(s). The system is language
agnostic so it is not simply returning other documents that use the same words;
rather it is returning documents regardless of how dissimilar they might appear
based on key word analysis alone. The
use of predictive coding was first approved by a court in Da Silva Moore v. Publicas Groupe SA, 2012 WL 607412 (S.D.N.Y. Feb.
24, 2012); since that time it has become the subject of comment (and
controversy) in the eDiscovery community.
The ultimate benefit of
predictive coding, and other advanced machine learning techniques, is not
simply the decrease in labor hours (and subsequent cost to your client). Instead, the ultimate benefit is to
extrapolate the lawyer’s knowledge and insight to data volumes once thought
unimaginable. This extrapolation results
in the ability to rapidly find what actually matters in the case, and leverage
that information in a strategic manner.
In the construction context it can
be vitally important for a party trying to prove (and even disprove) a claim of
impossibility of performance by providing deep insight into the digital
conversations of key players in the case.
In several instances, the party claiming unforeseen circumstances made
performance impossible has produced documents – without the aid of this
technology – only to have those documents examined using advanced technologies
such as predictive coding, and having their entire theory undone by the
data. Predictive coding can isolate the
key documents that showed what the performing party knew at the time of
contract formation, and more importantly what they did to explore alternative performance. The data often stands in stark contrast to
the claims made in the pleadings, and it becomes clear what the performing
party really knew (and when they knew it).
Predictive coding, as part of a system of search
techniques provides deep insight into what actually happened once performance
became impracticable, but does so in a fraction of the time generally
associated with the review process. Parties
on either side of the line in construction cases are well advised to discuss
the use of this technology with their clients.