1. Home
  2. Insights
  3. Articles
  4. Myths and Facts about Technology Assisted Review


Myths and facts about technology-assisted review

As technology becomes further integrated into our daily lives, technology-assisted review (TAR) is establishing itself into standard e-discovery practices. Also referred to as computer-assisted review or predictive coding, TAR is still fairly new. Five years ago, TAR was not even heard of in litigation, but today it’s considered an essential tool in e-discovery with more than half of U.S. corporations reporting using TAR in e-discovery last year. The demand for TAR continues to grow, driven by rapid expansion of large quantities of data tied to litigation and the rising cost to review that data. TAR software can save significant amounts of time and money during the review process. However, there’s still confusion and myths about what it can and can’t be used for. We’ll set the record straight and address the myths and facts about TAR.

Myth: TAR is highly complex

 While the backend and statistical programming portion of TAR may be complex, the basic premise of how it works is fairly simple. Reviewers use a set of documents, known as a seed set, to ‘train’ the TAR software. Reviewers code the documents as responsive or unresponsive based on their relevance and inputs into the predictive coding software. During the training, reviewers will verify if software is correctly identifying documents, and input more coding as needed. The software analyzes the coded documents and creates its own algorithms for reviewing and identifying document responsiveness.

Myth: Humans are better at reviewing documents than TAR

Fact: People are reluctant to trust technology to do a better job than humans when it comes to reviewing documents tied to litigation, despite studies proving otherwise. A 1985 study by David C. Blair and M.E. Maron found that TAR software outperformed human review in accurately identifying responsive documents. Attorney-supervised paralegals believed they had found at least 75% of relevant documents, using search terms and iterative search, but had actually only found 20% while TAR identified 75% of the responsive documents. This demonstrates that while the perception that humans are better at reviewing documents than technology persists, it’s not proved to be true. As TAR becomes widely adopted into daily practices and continues to prove its accuracy, people are likely to put more trust in it as a reliable tool.

Myth: TAR can replace human review

Fact: While the idea that technology could replace humans makes for a dramatic plot twist, it’s an unrealistic scenario. Human review and software programming are vital to the success of TAR. Categorizing documents as responsive or unresponsive does not happen automatically. TAR requires senior attorneys and SME to influence how the software determines the responsiveness of documents. The technology is augmenting our own human abilities, detecting seemingly invisible patterns as they review documents at speeds impossible for humans. TAR supports human reviewers, who now only need to review a fraction of the documents collected as opposed to the entire collection, allowing humans to make better and more accurate decisions.

Myth: Courts don’t support TAR

Fact: Today, courts widely accept the use of TAR in the e-discovery process. The first case to uphold the use of TAR in e-discovery was Federal Magistrate Judge Andrew Peck's decision in Da Silva Moore v. Publicis Groupe (Southern District of New York, 2012). He determined that TAR is the best and most efficient tool in most cases.

Key factors in Peck’s decision to endorse TAR included the vast amount of documents able to be reviewed, superior performance over other tools, cost effectiveness, and the transparent process. Peck reinforced the importance of transparency as he stated, “Such transparency allows the opposing counsel (and the Court) to be more comfortable with computer-assisted review, reducing fears about the so-called ‘black box’ of the technology. This Court highly recommends that counsel in future cases be willing to at least discuss, if not agree to, such transparency in the computer-assisted review process.” It can be assumed that as long as litigators consider these key factors in their own TAR process, we can expect to see more cases supported by this efficient tool.

Myth: TAR is 100% accurate

Fact: While TAR is not error free, that’s okay because neither is human review. What is more important to consider is that technology-assisted review is more efficient and accurate than human review alone. The purpose of TAR is to provide humans with a smaller, more relevant set of documents to review. When humans look at a smaller set of documents, they’re less likely to make errors, resulting in higher accuracy on the human side of review as well. It is important to consider quality control when using TAR in order to maintain the accuracy of identifying relevant documents. Documents identified as non-responsive should also be reviewed and searched for keywords. If these documents are found to be non-responsive, you can trust the quality of your TAR software.

Myth: TAR can review any type of document

Fact: Unfortunately, just because a document is stored electronically, doesn’t automatically mean it can be reviewed by technology software. TAR can only review text-rich documents such as emails, Word documents, PowerPoints, etc. TAR cannot be used to review spreadsheets, drawing files, images, blueprints, or video. This is because TAR programing uses text to code and identify responsive documents. If there’s no text information in the document, there’s nothing for the software to identify. That’s why it’s important to have a comprehensive strategy that incorporates a plan to review documents that are not suitable for TAR in your review process.

How will TAR evolve next?

As the amount of electronic information continues to grow, we can expect the same of technology-assisted review in e-discovery. Supported by courts, as well as litigators, for its ability to process enormous amounts of data and cut costs, with superior accuracy than human reviewers, TAR is becoming more established in the e-discovery process. As it continues to prove to be an invaluable tool in litigation, producing faster, more accurate results, we can expect some of the myths about TAR to disappear and people to put more trust in technology assisted review.