My, What a Difference a Year Makes

Disclaimer (A bit tongue and cheek, but I do think this is necessary as there are some critiques below):  The views and opinions expressed herein are solely those of the author and should not be attributed to Counsel On Call, Inc., or any of its officers, attorneys or employees.

I’ve had the good fortune to attend the past two annual conferences for the Association of Certified E-Discovery Specialists (ACEDS) in Hollywood, Fla., held at the superb Westin Diplomat.  In comparing the two conferences, all I can say is, “What a difference a year makes.”

First, and purely incidentally, the weather in 2012 was sunny, warm and generally quite pleasant. This year, the weather was overcast, rainy and a bit cooler. 

Just as the weather cooled a bit I think I detected a slight “cooling” of the conference attendees’ collective enchantment with the so called predictive coding technology.  In 2012, predictive coding was going to cause an industry and professional upheaval, eliminating the need for discovery (contract) attorneys, cutting costs, improving accuracy and possibly shifting influence between different stakeholders in this area.  One year later, we have experience  – more reported court decisions directly on point and more vendor entries into the marketplace.  With this collective experience, the conference attendees had a cooler, more nuanced view of the technology.

Please don’t misunderstand, there’s no question that predictive coding (also known somewhat synonymously as technology-assisted review or simply “TAR”) is here to stay and should only improve with time.  Rather, the bloom has worn off, and practitioners are discovering that, although at times and for certain types of matters, TAR improves efficiency, overall quality of a review and can significantly lower overall costs. Nevertheless, it is neither a cure-all nor the disruptive technology that some claimed last year.

I think there are several reasons for this maturing of the collective view:

  • The term predictive coding (trademark issues aside) seems to mean different things to different people, hence the use of the more generic TAR designation. This causes confusion among potential customers.
  • Some technology vendors may have rebranded older technology as TAR, perhaps thereby lessening the user experience.
  • Different TAR tools have different “blind spots” that limit their utility, e.g., image files and spreadsheets may not be considered by the analytics.
  • Far from removing human judgment from the process, TAR applications may increase dependence on human judgment.  For example, mistakes by “subject matter experts” can be amplified.  Alternately, I suggest trying your hand at picking “exemplar” documents to teach the computer – a document might be technically non-responsive to the litigation but would nevertheless have excellent teaching parameters.
  • TAR itself is not inexpensive.
  • TAR reduces data but does not eliminate the need for some linear review, either with quality control or during the construction of a privilege log.
  • Not all matters are suitable for TAR, ether due to the size of the case or the type of data.

Counsel On Call has always been a solid proponent of predictive coding as well as an early adopter.  Vendors now, however, call their technology predictive coding without the functionality.  There is no question that predictive coding is here to stay; rather, its potential is still less than hyped.

I’m looking forward to attending next year’s ACEDS Conference.  While not perfect (panels are too large and too much time is spent on speaker introductions), it’s the only conference of which I’m aware that is focused on the e-discovery practitioner.  ACEDS also seeks to professionalize this field, and this is a good thing.

Maybe next year the weather will be warmer.  It will be interesting to take the attendees’ temperature on TAR as well.


What Would Hal 9000 Do (WWH9D) for E-Discovery?

“Greetings, attorney 61432. Your security access has been confirmed. What type of documents would you like me to search and code for you today?”

Imagine e-discovery being that easy. Simply logging on and telling a computer to handle your process for you. Sounds like something out of “2001: A Space Odyssey.”

When I began my legal career, the dominant discovery work involved reviewing paper documents in cardboard boxes and inputting information onto coding sheets or onto computer ‘boxes’ with blinking orange or green lights and no internal drives. Not exactly the golden age of artificial intelligence, eh? That being said, very few could foresee the impact of the computer itself in the legal world, but with the advent of e-mail an entire industry was born to deal with the “E” part of e-discovery in particular. The jobs that many lawyers hold today did not even exist in that form just a decade ago.

Fast forward to today where the talk of the town is predictive coding using the ones and zeros to do the work that a human lawyer currently does (maybe "The Matrix" would've been a better analogy here). Predictive coding has received an increasing amount of discussion among the various forward-thinking lawyers involved in e-discovery (see a good example here). This trend and its impact will no doubt increase over time.

Predictive coding takes a number of forms but basically it involves someone with case knowledge starting to review documents, and then the computer learns from the documents marked responsive, non-responsive, privileged, etc., and applies that knowledge to categorize all the documents in the entire data set that has been collected. (Of course, the value that most discovery attorneys bring to a case is more than just tagging a document as responsive or privileged. They also analyze the documents and determine how they fit into the overall litigation. They can find problematic, hot or helpful documents too that might not be possible to find strictly from a predictive coding perspective.)

The question that remains is whether predictive coding will be good enough to rely on for production or will we still need to review the documents at a more “Quality Control level” before production. The upsides are definitely there if it can work. Speed, cost savings, accuracy, etc., could all be improved. As always, much depends on the accuracy of the initial information inputted into the process, but overall it bodes well for saving clients money.

And as cost-saving devices are created to meet the current technology challenges, the technology challenges themselves are also changing and increasing. The next challenges will likely involve social media sites, text messaging, audio and video files, and dynamic databases, each of which are only starting to rear their ugly heads in the discovery world.

Add all of this discussion to our current economic times and we once again have the concern about the impact on jobs and future careers. Will they still need me if a computer does all the review? The question is not unique to the legal world and the answer may surprise some of you.

While it took a decade for e-discovery to take over the entire discovery process and discussion, the new waves will likely come much faster as technology evolves and impacts everything that lawyers, especially attorneys involved in e-discovery, do. Attorneys may not be first-level reviewing any e-mails in five or ten years, but they could be looking at videos on an eighth-generation iPad! New avenues that currently don’t exist or only exist in their embryonic stage will be available to us all. There will certainly be new workflows and protocols created to incorporate technology into making the discovery process more effective and efficient – someone is always seeking ways to build a better mousetrap, and we at Counsel On Call fall into that category.

Part of my job as Discovery Process Architect is to continually analyze what’s working and what isn’t, what our teams can do to better utilize the available resources, how to deliver the best-possible work product, and ultimately what can we do to save our clients time and money. Predictive coding could very well be a part of that in the not-too-distant future. I do believe, however, that attorneys will still be an integral part of any smart and flexible discovery processes that are developed.

The bottom line: I don’t believe we’re anywhere near the Hal 9000/Space Odyssey scenario or a comprehensive predictive coding process eliminating the use of attorneys during the review.

But as the year nears its close, realize that change is constantly in the air these days. Change with it. Expand your skill set. Don’t get stuck where you are. Learn something new about technology this year. Read. Experiment. Become better at what you do.

I know these words have been uttered by every generation, but as one of the old guard who has transitioned into our brave new e-discovery world, I can speak to the value of its counsel.