The NYTimes had an article this weekend about the growing number of e-discovery vendors who can go beyond mere keyword searches into linguistic and sociological reasoning about millions of pages of documents:

[T]hanks to advances in artificial intelligence, “e-discovery” software can analyze documents in a fraction of the time for a fraction of the cost. In January, for example, Blackstone Discovery of Palo Alto, Calif., helped analyze 1.5 million documents for less than $100,000.

Some programs go beyond just finding documents with relevant terms at computer speeds. They can extract relevant concepts — like documents relevant to social protest in the Middle East — even in the absence of specific terms, and deduce patterns of behavior that would have eluded lawyers examining millions of documents.

It often comes as a surprise to non-lawyers and law students, but the bulk of litigation work (measured in hours) performed at big law firms doesn’t really involve legal training or legal reasoning. The bulk of the work — which is performed by junior associates and contract attorneys as part of the “leverage” business model — involves the dreaded “document review.”

 



The Old World Of Document Review

Let’s say that you’re a medical device company that just recalled nearly 100,000 hip replacement implants because the darn things were never tested and have an unfortunate tendency to release metal shavings into their patients’ surrounding tissue and blood stream. (I’m looking at you, DePuy.) You have already been sued by a couple thousand of those patients, and are likely going to get sued by virtually all of them.

In the course of those lawsuits, the patients are going to request, and you are going to be required to produce, all relevant documents that describe the testing performed on the devices, their propensity to fail, what you said to the FDA during the approval process, etc., etc. But you don’t want to produce those documents, since they know they will probably show that you were indeed negligent in the design, testing, and manufacture of the hip implants.

So what do you do? You try to claim attorney-client privilege for as many documents as you possibly can. To do that, however, you have to have at least some basis for claiming the privilege for each claimed document. You also try to claim that many of these documents are “irrelevant,” even when they, of course, are relevant. To do that, however, you have to have some basis for arguing why each claimed document is irrelevant.

It’s a time-intensive process. For a surprisingly long time, large companies that had done something very wrong were willing to pay hundreds of dollars an hour for recent law graduates in windowless offices tucked within skyscrapers to do this “document review.” The business model was already falling apart over the past few years, as big law firms realized they could charge almost as much for temporary, non-career-track contract attorneys (as compared to full-time, career-track associates) to do the same work, even though the contract attorneys cost half or less as much as associates do. Similarly, a number of corporate clients realized that they could outsource a lot of this drudgery work to places like India, where a good chunk of the population can read basic English and identify the sender and subject of emails just as well as most American lawyers.

The Rise Of Sociological And Linguistic E-Discovery Document Processing Software

The New York Times article, in essence, describes a couple software vendors who have come up with “linguistic” and “sociological” ways of reviewing documents, so that far more information can be extracted, and far more subtle relationships can be identified, than through a mere keyword search.

There’s no doubt that this next generation of document review is the way the future. It will be standard practice for companies and their attorneys to use the software to identify inculpatory and embarrassing materials so they can focus their efforts on precluding discovery into those materials.

But there is some cause for concern. Consider this paragraph in the story:

Software is also making its way into tasks that were the exclusive province of human decision makers, like loan and mortgage officers and tax accountants.

You mean, the dreaded robo-signers? The computer programs used by banks to routinely rip off millions of homeowners, including, most recently, a columnist for the Washington Post?

That’s supposed to be an indication of how well these programs work?

The bigger problem, in my view, is the asymmetrical manner in which these programs will be used. Again, the story provides a perfect example:

Two months ago, Autonomy, an e-discovery company based in Britain, worked with defense lawyers in a lawsuit brought against a large oil and gas company. The plaintiffs showed up during a pretrial negotiation with a list of words intended to be used to help select documents for use in the lawsuit.

“The plaintiffs asked for 500 keywords to search on,” said Mike Sullivan, chief executive of Autonomy Protect, the company’s e-discovery division.

In response, he said, the defense lawyers used those words to analyze their own documents during the negotiations, and those results helped them bargain more effectively, Mr. Sullivan said.

The whole purpose of discovery is to eliminate “trial by surprise.” The lawyers and the parties are supposed to have ample opportunity to investigate the facts and circumstances of the case before they settle or go to a jury trial. Yet, as shown by the above example, one-sided use of these programs can give the defense attorneys an unfair vantage by allowing them to perform a more robust analysis of the exact same documents.

There’s an easy solution, though. In a number of cases these days, the lawyers agree to a streamlined form of discovery in which the bulk of the documents are initially produced and, if privileged or irrelevant documents are discovered in the mix, then the party producing the documents reserves their right to "clawback" the documents and claim a privilege.

It works. It’s fair. It ensures that cases will be tried on their merits, not on the ability of lawyers to conceal evidence. But will that become standard practice?

I doubt it. There is too much money to be made in, and too much inculpatory evidence to hide with, obstructionism.
 

  • Dolin Duval

    Yes! It is a very exciting time for AI in eDiscovery. Not sure I would take on a large discovery without it today. The best I have seen is PureDiscovery out of Dallas. Don’t think anyone can touch them on scale. Their concept search engine is great.

  • Joe Thorpe

    After nearly 10 years of ai evolution in the field of ediscovery (exclusive domain of defense bar), Trial lawyers are finally starting to get it. See another great blog post by Doug Forrest on the subject: http://www.ilsteam.com/2011/08/19/ediscovery-is-the-age-of-the-cyborg-is-at-hand/