Is predIctIve codIng
the answer to
REDUCING THE COSTS
OF EDISCOVERY?
deborah Baron from nuix believes predictive coding must be
combined with other technologies and investigative workflows to
address the runaway costs of legal review.
the legal industry in recent years has looked to a machine-learning
technology known as “predictive coding” to make litigation more
affordable by reducing the number of documents that must be
reviewed by human beings before producing evidence to a court.
however, legal teams have traditionally used predictive coding only at
the final stage of discovery. In addition, predictive coding is only one
weapon in a potential arsenal of analytical tools and techniques that
can make electronic discovery less burdensome.
A GROWING BURDEN
The burden of discovery is increasing for organizations
facing litigation, regulatory disputes and audits. Surveys
of corporate counsel and executives show they expect to
face greater litigation and regulatory scrutiny. 1 Discovery
production obligations remain fixed, but the volume of
data involved grows each year.
Technology has made it easy to store and retrieve millions
of email messages or documents in seconds. However,
the cost for document review becomes expensive when a
human being has to review each page for responsiveness
and redact privileged information.
An often quoted study by the RAND Corporation found
manual document review accounted for almost three-quarters of eDiscovery production costs. 2 It estimated
review costs were an average $18,000 per gigabyte of data
and up to $30,000 per gigabytes in some circumstances.
To put this in context, a single high-end smartphone or
tablet device today has up to 64 gigabytes of internal
storage, while many personal computers have a terabyte
or more of disk storage.
