Practicalities of Computer Assisted Review

  • Blog Post
  • Posted on 22 July 2014

What do Google+, the new coke, iSnack 2.0 (the new Vegemite product before it was renamed to Cheesybite) and Assisted Review, also known as predictive coding, have in common? The answer, unfortunately, for Assisted Review is, all have had a lot of publicity about how they will change our lives, but no matter how good the product is none has had a huge uptake.

Opinion Piece: Written by Sam Lamble, Solutions Consultant at Law In Order

What do Google+, the new coke, iSnack 2.0 (the new Vegemite product before it was renamed to Cheesybite) and Assisted Review, also known as predictive coding, have in common? The answer, unfortunately, for Assisted Review is, all have had a lot of publicity about how they will change our lives, but no matter how good the product is none has had a huge uptake.

Lawyers aren’t mathematicians and don’t gamble with their clients or cases. So when you start throwing around terms like; 95% confidence interval; 2.5% margin of error; or Latent Semantic Indexing, eyes glaze over, risk sensors starting pulsing, defences are put up and the legal technology person suggesting these options is put into the category of ‘Techies that over-complicate the situation and never provide a straight answer’.

While the underlying technology is no doubt important in ensuring that Assisted Review is a success (and if you care to go into the detail, many articles have been published on this and there are many different software options available), the principle of ‘why’ is often the part worst explained in the whole process.

Humans make mistakes, even lawyers make mistakes, and ever since the dawn of electronic review, computers and software have assisted in identifying these mistakes and allowed them to be corrected quickly and easily. Anyone who has completed an electronic review, I’m sure, will admit that a quick search across the database after the initial review has been completed will return a handful of documents that have been marked as relevant but don’t have a designation for privilege, or should be redacted but were skipped in your mad panic to meet a deadline. In these situations, you would find it hard to imagine how a legal review process was undertaken in a time before electronic review databases like Relativity or Ringtail and the relative worth of such systems is immediately clear.

Assisted Review is simply an extension of this lawyer and computer relationship, but rather than waiting until the end for a lawyer to let the computer assist, the computer learns from lawyer/s as they work. Likewise, a lawyer will learn from the computer and challenge any wrong ideas as they arise. Only once the computer and lawyer reach a level of confidence with each other the value of Assisted Review is appreciated.

Having recently completed a matter using Relativity’s Assisted Review integration with some really great results, I can tell you that it was amazing to have the lawyer and legal technology experts Andrew King (Litigation Support Consultant, eDiscovery Consulting NZ), Martin Flavell (eDiscovery Manager, Law In Order) and I celebrated the overwhelming cost and time saving for the client using the Computer Assisted Review Software. It was even more amazing that we all were confident with the software before it was confident with us. For more details on this, Martin recently completed a case study, which can be found here.

Share this post