Читать книгу Insiders' Guide to Technology-Assisted Review (TAR) - Коллектив авторов - Страница 4

The evolutionary path of modern Technology-Assisted Review (TAR) was paved as much by necessity as it was by innovation. These advanced solutions are now necessary because traditional review workflows simply cannot keep pace with the growing volume of electronically stored information (ESI). Moreover, heightened awareness of the potential flaws associated with linear review1 has called unwanted attention to the very real potential for inaccurate coding and incomplete productions.

At the same time, the underlying discipline that drives TAR methodology – the art and science of information retrieval (IR) – has been widely accepted in government and industry-specific sectors (accounting, finance, insurance, telecommunications, and health care to name a few). This active and ongoing use of IR methodology to search, mine, and manage large sets of electronic data has allowed these innovative solutions to go through decades of testing and validation before reaching the mainstream legal marketplace. And while the use of technology is not new or novel in the discovery arena, the road to modern TAR has been paved with numerous iterations of technology-assisted workflows.

In the literal sense, “technology-assisted review” has been ongoing since the first computer was used to help log or categorize a set of documents in response to a request for production. Database capabilities and sophistication grew rapidly during this inaugural wave of technology-assisted solutions, allowing for document collections and productions to be tracked through delimited fields and warehoused indefinitely.

Further advancements came with the ability to scan the physical image of each document and render that image as a static file in a separate database that could be accessed on demand by reviewers through network connections. Combined with what was then considered a cutting-edge technology called optical character recognition (OCR), reviewers could access search results for both the image of a document and its underlying text in one location. Technology-assisted review was in its renaissance and the legal community fully embraced these solutions.

This was the dawn of the digital age in which paper collections were starting to be replaced by word processing tools and the default standard for business correspondence switched from written letters to electronically transmitted e-mails. With this sea change under way, yet another wave of literal “technology-assisted review” solutions was under development: data processing and hosting applications.

Electronic documents must go through a series of highly technical steps before they can be reviewed. The data needs to be collected (without corrupting or spoliating the evidence), standardized to isolate only the user-created files, deduplicated to reduce the cost of reviewing the same document twice, and then rendered for review in a hosted application that presents a genuine version of the original native file for analysis. Breakdowns during any of these steps will result in the loss of important information about the document and the potential that the data will be inadmissible if presented as evidence in a court of law.

One of the greatest advantages of data processing is that we learn a lot about the data set as it is staged for review. The extracted text and metadata for each document contain troves of information that can be used to help organize these files in a meaningful way. That effort can be as simple as grouping the documents by custodian or by search results, or made more complex by introducing some of the early TAR systems such as clustering or ranking the documents based on underlying content. These solutions are the natural progression of performing legal discovery in the digital age and represent the starting point for modern TAR in application.

TAR solutions, in all of the structural varieties outlined in Chapter 1, as well as during their early developmental states described above, still need to be executed through a workflow that involves trained professionals reviewing a collection of documents. Therefore, it is equally important to consider the review process alongside evolving technology to fully understand the evolutionary path.

From the early 1980s to mid-1990s, legal document reviews were fairly straightforward: locate just enough space to accommodate any given set of boxes and the appropriate number of people needed to review the contents, fill the seats with junior lawyers or senior paralegals, and start the project. Reviews were exclusively linear in this time period; the contents of each box would be reviewed from front to back with some color coding or physical marking to identify relevant and/or privileged documents. Little if any thought was given to technology, infrastructure, or workspace ergonomics and only modest consideration was paid to the workforce supporting this effort – reviews were often performed in windowless basements or off-site archive facilities with limited amenities, effectively isolated from the legal case team.

This underscores the fundamental problem with discovery in the digital age: it has become disconnected from the legal process that gave rise to the obligation in the first place. The ultimate goal of discovery is to facilitate a complete and accurate factual record and to put the salient facts on the table so practitioners can advocate for the best interests of their clients based on the merits of the case.

Document review in the new millennium has responded to many of these concerns. Review centers have been moved out of the basement into fully modernized facilities with workspace designed to ensure the security of client data as well as reviewer comfort. Centers are equipped with redundant high-speed internet connections and a dedicated IT staff to support the technology infrastructure that is the backbone of a modern document review. There is an enormous pool of talented and experienced review attorneys with specialized substantive skills to meet the needs of any legal matter. Finally, workflows are being designed and implemented by experienced project managers to ensure that efficiencies are realized and costs are reduced wherever possible.

Despite the many advances in early TAR solutions and document review protocols, the ability to accurately and consistently analyze data is simply not able to keep pace with the proliferation of ESI in today's corporate and social environments. There are simply too many data sources and too much raw data. So how can the legal community feel confident that modern TAR methods are achieving their intended purpose? How can modern TAR reconnect discovery with the merits of the case, reduce cost, and improve quality?

The answer to that lies in the chapters ahead and detailed discussions of the next generation of TAR solutions, increased judicial acceptance, and the application of modern TAR to legal review above and beyond discovery.