Predictive coding technology the use of computer software to assist and expedite discovery document review is the latest e-discovery tool with which litigation counsel must become familiar. By potentially eliminating the need for large-scale document review, predictive coding technology can reduce costs and streamline the e-discovery process. Proponents also argue that computer-assisted review can be more accurate than traditional attorney review. However, as with any new technology, counsel must be prepared to assess its usefulness in a particular case and to address potential client concerns.
Parties implementing predictive coding: (i) start with a set of data, derived from “traditional” methods (e.g., keyword searching); (ii) review a “seed” or sample set of documents for responsiveness and/or privilege; (iii) employ computer-learning software to review and categorize potentially responsive documents; and (iv) analyze the results to obtain quality-control feedback and achieve coding consistency. This process is repeated until the party is confident that the results of the computer-assisted review gibe with the responsiveness and privilege determinations made through counsel’s review of the seed set.
