Predictive Coding Could Doom Document Review by Humans

As if lawyers need any more reasons to fear for their livelihoods in a tight job market, here's another: Predictive coding is becoming more popular, and could make human document review somewhat obsolete.

Yes, the rise of robots is reaching into the realm of entry-level and contract legal work, the website Lawyerist reports. What used to take a roomful of document reviewers days, weeks, or even months to complete, can now largely be done by computer programs designed to quickly sift through terabytes of electronic data.

Experts say predictive coding saves time, money, and cuts down on human error. Human document review isn't totally doomed, they say, but document-review lawyers should still prepare for changes linked to law firm automation.

Though predictive coding is computerized, it's still a lawyer-driven process. Attorneys define search parameters for the predictive coding algorithm, which then digs for the most relevant documents.

Attorneys can repeat the process several times by refining their searches and getting more precise results.

Some practitioners have high praise for predictive coding. It's often "superior" to an initial manual review of documents, according to a recent study in the Richmond Journal of Law and Technology.

But the trend doesn't totally spell trouble for document review attorneys. After predictive coding identifies the most relevant documents, a human is still needed to review those documents for responsiveness to a discovery request, and for any privileges that may keep the documents out of play.

For document reviewers who are still concerned for their jobs, one writer for Lawyerist offered these recommendations, among others:

• Network during your document review assignment. It could lead to new document-review contracts, or even full-time employment.
• Learn to do more in-depth document reviews. Your ability to perform second-level or quality-control reviews can make you invaluable to a firm. Higher-level reviews are also tasks that predictive coding currently can't touch.

Keeping these ideas in mind just might slow, if not stop the rise of the machine.

Related Resources:

• E-Discovery And The Rise of Predictive Coding (Forbes)
• Imaging/Scanning/Coding - Practice Support & Consultants (FindLaw)
• Legal Technology Videos - Document Review (FindLaw)
• Evolution of Document Review (FindLaw)