Fun With Electronic Redaction

In eDiscovery, redaction is the process of blocking or removing sensitive information from reviewed documents. This doesn't mean just simple viewing controls, but also protecting privileged information against copy, search or print.  Why not just suppress the entire document? Because if certain information in a document is relevant then you must include it in the document set, and you must protect any privileged information within by other means. The ability to do this gets a lot of press around eDiscovery but it also applies to compliance, Freedom of Information Act, and so on - any time documents are reasonably demanded but a portion of the information must be protected.

In the good old days of simple paper production, redaction was simple. The reviewer read the documents with a black marker in hand and slashed away. Now redaction is a lot -- a lot -- harder. The vast majority of reviewed data is electronic. Although some smaller firms have been known to print everything out and redact the printouts, this is a clumsy operation to say the least. (And of course entirely unworkable for any eDiscovery matter beyond the tiniest case.) Even when the production consists of imaged TIFF files with printed redaction blocks, reviewers still need to apply those redactions on-screen to the electronic documents.  

Redaction gets much more complicated when reviewers are dealing with native application formats. Sure it sounds easy to redact a Word file. Just select the privileged block and add a black highlight. Or for the more sophisticated Word user, hide the block against viewing. The problem with both of these approaches is that when opposing counsel views them on native applications, then it is the simplest job in the world to copy and paste it into a new document to reveal the presumably privileged text. Think this doesn't happen? Check with Schaefer v. GE, where the plaintiff's redaction mistake resulted in potential threats to both plaintiff and mighty GE.

There are redaction technologies out there to help and I strongly urge companies to look into them. Baseline features for such a product should include: 1) automatic redaction, 2) on-screen manual redaction, 3) defensible redaction choices, and 4) preservable redaction marks.

Automatic redaction should include search-based redaction and pattern matching redaction. Redaction should protect against view, print, copy and search.On-screen manual redaction provides for on-screen redaction. As with automatic redaction, it should protect against any unwanted intrusion into redacted content.

Defensible redaction enables tagging defensible reasons for the redaction (i.e. "Privileged"), and also defends against spoliation when using native applications. The software redacts copies of the source file to avoid modifying the original.

Preserving redaction marks should operate throughout the production process. It sounds like it should be obvious but it isn't - not all production platforms will retain redaction marks from during the review process. Make sure your software will.