Liz Eager's Article on Predictive Coding Software Featured in the Fulton County Daily Report
Liz Eager's article on the pros and cons of using advanced technology for document reivew is featured in the Fulton County Daily Report. Below is the full text of the article.
Last year I settled a case in which we had to review more than 30 million pages of documents. The majority of these documents were likely not relevant or duplicates, yet we still faced the daunting prospect of dealing with these documents. Given the overwhelming volume of electronic communications, it is not uncommon to see cases with such vast quantities of information.
While technology is largely responsible for creating this problem, technology may also help us solve it. New software is being developed and used with the objective not only to reduce the time and cost of discovery but also to improve the overall quality of discovery, meaning it increases the chances of capturing all of the relevant, responsive documents with the smallest number of non-responsive ones.
These new technologies are referred to as predictive coding or concept searching software. These programs claim to "learn" the set of documents by keeping track of the documents the reviewer tags as responsive or relevant, analyzing the information in those documents and pointing the reviewer to similar documents in the collection. If the software can successfully achieve this objective, it could change the way discovery is currently handled.
One potential benefit of this technology is its use as a tool to perform a more thorough early case assessment so that you can analyze a case's strengths and weaknesses. The technology's advanced search features can quickly cull through documents and make sure that you are not taking positions inconsistent with your documents. It also can get rid of spam and other worthless documents leaving you with fewer documents to review, thus reducing the overall time and cost of discovery.
But there are also concerns with using this type of software that you need to consider. There is the potential that the software may miss important documents. In some cases the most damning documents are simple emails saying "we are going to take it on the nose for this one" or "so that's it, right." The software may not locate these types of documents.
It may also miss documents where code words are used to describe transactions at issue in the case or documents intentionally worded to be misleading. Another potential issue is that you may miss concepts not considered at the outset of a case. Sometimes the documents tell a different story than what was originally thought, and your case strategy may shift. But if you did not originally "teach" the software to capture documents related to the new case theory, the software may have weeded out those important documents. By eliminating these documents, the software may miss the documents that make—or break—your case.
All of this being said, the potential for this software is tremendous, but it should never be solely relied upon and lawyers should always take steps to ensure they don't miss the smoking gun. It also will work better in certain cases and not as well in others. It is important for those who are using it to identify which cases are good candidates for this new technology and which cases are not.