The dire “prediction” continues in the blogosphere, newspapers and law firms that predictive coding will somehow eliminate the need for human review of documents during EDiscovery, and eliminate key legal jobs. The amount of world information (data) continues to double every two years, which makes it increasingly difficult for company’s to organize, track and identify information as responsive to EDiscovery requests. I believe that it is wishful thinking on the part of corporate law departments who are looking at the high cost of historical document reviw and hoping that technology will provide a low cost answer to the review of discovery documents.
The “history of discovery can easily be summarized in a few key steps:
• It wasn’t that many years ago when most discovery requests involved the exchange of only a few dozen boxes of information between attorneys. This information was reviewed by the lead counsel and a few associates over the course of a couple of weeks with certain documents identified as relevant for litigation. Oh, the good old days!
• Next came the discovery was the practice of “dumping” hundreds of boxes of documents on the opposition, burying the key documents throughout the boxes making it difficult to locate them. This is when the rise of linear review by teams of attorneys and paralegals became commonplace. It was the only practical way to get through the documents, and hopefully identify the key documents. Generally, you had one chance to locate documents.
• .Soon, databases were created that contained document summaries of key “fielded information” that allowed the attorneys to search through the database multiple times, using the information in the fields. The creation of a fielded database of every document was extremely expensive and time consuming and was often filled with errors due to the differences in review and legal experience of the review teams.
• For the last 10 -12 years, early EDiscovery emerged due to the rise of computerization and the increase in data. EDiscovery allowed us to perform online review of the total data set, which was a vast improvement over the handling of documents or document summaries. But the process was still lengthy and expensive to implement.
• Currently, EDiscovery involves collecting, processing and hosting documents, emails and other document types. Advanced technology allowed us to de-dupe populations, exclude non-responsive documents, remove personal emails and exclude documents using date range searches. The time of complex Boolean search technology allowed us to reduce document populations and allow teams of lawyers to review them all before they deliver “responsive” data to the opposition.
• The volume of data that must be included in EDiscovery has recently grown to include personal laptops, Smartphone’s, the cloud, social media, etc. Along with this increase in the volume of data, EDiscovery vendors have created advanced technology to help parse through the data. The emergence of predictive tagging has brought the technology further to the forefront. Predictive tagging (or analysis) involves “training” the software using algorithms to identify categories of relevant documents for each case. The algorithms help the programs to “learn” to perform iterative passes through the data, identifying documents that match what the attorneys have identified as relevant documents. And then, once it has been created the subset of documents identified must still be reviewed by a team of attorneys to verify all the decisions made by those developing the algorithms. While it is amazing technology that goes a long way to reducing vast quantities of data to a more manageable document population size, it must be driven by extremely smart attorneys and technologists, or the system will fail and return garbage for results.
What is the common link throughout this short history of discovery? That attorney reviewers and human interaction have been key to the success of litigation from the start. At each phase of the process, the role of attorney reviewers has changed, but there can be no denying that even with the advanced technology that is emerging, attorney reviewers will continue to play a strong and central role. Without human guidance and verification, computers can not handle the complex decision making that ultimately determines whether a single document is responsive or relevant to a particular set of circumstances.
Predictive tagging requires that a team of lawyers knowledgeable about the case must first manually go through a sample set of documents and teach the program what is relevant and what is not. This is an iterative process that must be performed, refined and repeated in order to create a relevant search of the total population. Once this baseline of documents has been identified, legal teams can benefit from the software’s quick prioritization of data. But ultimately, it is still the lawyer’s job to confirm the results and add new search parameters based on each set of results. This also involves complex use of sampling techniques and iterative processing.
The Takeaway
The review of data using advanced technology highlights the collaboration that must exist between technology and human review. Automation technology is the engine that improves the process, but so far, it can not drive it without people. It is clear that people are an indispensible part of the success of implementing technology into document review. While the technology can speed the process and dramatically reduce the number of documents which ultimately need to be reviewed, human involvement is key to both the success of the culling and the final review of the documents identified as fitting the search parameters.
Software is just a tool to be used effectively by lawyers to help identify relevant information in a sea of data. But people need to remember that it is the human mind that can identify a parallel course of investigation from reviewing the content of a document “selected” by using comprehension to process nuance in language, intuition and logic. Critical thinking will always be the most important process involved in the discovery process. Document review processes will change over time. The number of reviewers needed to perform this work is likely to decline over time, but the effectiveness of review teams will require lawyers and review teams to have a higher level of technological and legal background to be effective at this very important job. Those companies that harness this power and create processes to exploit the new technology while employing legal staff to perform a more sophisticated type of document review will rise to the top.
Comments