Aug 6 (Reuters) - More than 90 percent of a small group of
in-house counsel and law firm attorneys said the computer coding
used to sift through electronic documents was useful, though it
might not be the big money saver it is sometimes billed as, a
new survey reported.
Thirteen in-house counsel from Fortune 1000 companies and
11 Am Law 200 partners and senior counsel were asked about their
experiences with so-called predictive coding, the case-specific
searches used to predict whether documents are relevant to
litigation. The survey, conducted on behalf of FTI Consulting,
which manages e-discovery for large corporations, did not
identify its participants by name but found that 54 percent had
used the technology to aid in discovery.
While many in the legal industry have worried that such
coding will replace human review, very few respondents said they
relied on the technology to automatically indicate whether
documents were relevant or not. Instead, they used it to cull
and prioritize documents, so the process would be easier for
lawyers reviewing the material, the report said. "With
predictive coding, you are getting your hands on the key
documents much earlier in the process than in more traditional
approaches," a law firm respondent told the surveyors, who
included legal marketing consultant Ari Kaplan and FTI's Joe
Most respondents said the algorithms sped up document
review, but they were unable to exactly quantify how much money
the technology saved them. Twenty-seven percent could not assess
the costs; the same percentage estimated cost savings between
$25,000 and 250,000. Another 18 percent estimated savings
between $250,000 and $500,000, while 27 percent believed they
saved more than $500,000. One of the respondents who believed
they had saved between $25,000 and $50,000 said the process
allowed them to avoid reviewing 80,000 to 90,000 documents.
The lawyers found predictive coding especially useful in big
cases involving more than 100,000 documents, cases requiring
documents to be gathered from 50 or more people, in commercial
litigation and when responding to governmental requests
requiring large-scale review in a short time period. The
technology wasn't so useful in employment cases, patent
litigation and cases where a large number of documents to be
reviewed are images or audio files rather than text.
Responses to the questions were largely consistent, the
study's authors told Reuters, though in-house attorneys were
more concerned with developments in how courts are addressing
and responding to the technology, Looby said.
Jeffrey Fowler, a partner at O'Melveny & Myers who is part
of his firm's electronic discovery practice, said that
predictive coding is necessary in e-discovery. But he wasn't
surprised that some companies expressed concern about judges'
attitudes toward the technology. Members of the judiciary have
sometimes asked litigating parties to work together and disclose
their processes to each another, Fowler said.
"Clients are concerned that if they use predictive coding
... (they) will have to involve the opposition at a level that
will be too costly and not practical."
(Reporting By Erin Geiger Smith)
Follow us on Twitter @ReutersLegal | Like us on Facebook